Skip to content

Latest commit

 

History

History
1865 lines (1504 loc) · 81.1 KB

v0.4.0.summary.md

File metadata and controls

1865 lines (1504 loc) · 81.1 KB

Table of Contents

  1. Overview
  2. Example Images
  3. Mixed-Category Patches
  4. Added
  5. Changed
  6. Refactored
  7. Fixed

Overview

Release 0.4.0 focused mainly on adding new augmenters and improving the internal augmentation "backend".

The following augmenters were added (see the overview docs for more details):

  • ChangeColorTemperature: Gives images a red, orange or blue touch.
  • New Brightness augmenters: WithBrightnessChannels, MultiplyAndAddToBrightness, MultiplyBrightness, AddToBrightness.
  • New Dropout augmenters: Dropout2d, TotalDropout.
  • RemoveSaturation: Decreases the saturation of colors. Effects are similar to Grayscale.
  • Cartoon: Applies a cartoon-style to images (classical / not-learned).
  • MeanShiftBlur: Blurs images using a mean-shift clustering method. (Note: Very slow.)
  • Jigsaw: Splits the image into rectangular cells and randomly switches some pairs of neighbouring cells. (Note: Does not support bounding boxes, polygons and line strings.)
  • WithPolarWarping: Transforms images to polar coordinate space and applies child augmenters there.
  • SaveDebugImageEveryNBatches: Generates and saves at every N-th batch a debug image visualizing all inputs within the batch. Useful to gauge strength and effects of augmentations and quickly spot errors in ground truth data (e.g. misaligned bounding boxes).
  • Cutout: Removes rectangular subregions of images. Has some similarity with CoarseDropout.
  • Rain and RainLayer: Adds rain-like effects to images.
  • RandAugment: Combination of multiple augmenters. Similar to the paper description. (Note: Can currently only augment images.)
  • Identity: Same as Noop. Does nothing.
  • UniformColorQuantizationToNBits: Quantizes each image array component down to N bits. Similar to UniformColorQuantization. Has the alias Posterize.
  • Solarize: Invert with threshold.
  • RemoveCBAsByOutOfImageFraction, ClipCBAsToImagePlanes: Augmenters to remove or clip coordinate-based augmentables, e.g. bounding boxes
  • More blend augmenters:
    • BlendAlphaMask: Uses batch-wise generated masks for alpha-blending.
    • BlendAlphaSomeColors: Alpha-blends only within image regions having specific randomly chosen colors.
    • BlendAlphaSegMapClassIds: Alpha-blends only within image regions having specific class ids in segmentation maps.
    • BlendAlphaBoundingBoxes: Alpha-blends only within image regions covered by bounding boxes having specific labels.
    • BlendAlphaHorizontalLinearGradient: Alpha-blends using horizontal linear gradients.
    • BlendAlphaVerticalLinearGradient: Analogous.
    • BlendAlphaRegularGrid: Places a regular grid on each image and samples one alpha value per grid cell. Can be used e.g. to achieve coarse dropout.
    • BlendAlphaCheckerboard: Places also a regular grid on each image, but neighbouring cells use alpha values that are inverse to each other.
  • Shortcuts for Affine: ScaleX, ScaleY, TranslateX, TranslateY, Rotate, ShearX, ShearY.
  • Several new crop and pad augmenters: CenterCropToFixedSize, CenterPadToFixedSize, CropToMultiplesOf, CenterCropToMultiplesOf, PadToMultiplesOf, CenterPadToMultiplesOf, CropToPowersOf, CenterCropToPowersOf, PadToPowersOf, CenterPadToPowersOf, CropToAspectRatio, CenterCropToAspectRatio, PadToAspectRatio, CenterPadToAspectRatio, PadToSquare, CenterPadToSquare, CropToSquare, CenterCropToSquare.
  • Wrappers around imagecorruptions package (verified to have identical outputs): GaussianNoise, ShotNoise, ImpulseNoise, SpeckleNoise, GaussianBlur, GlassBlur, DefocusBlur, MotionBlur, ZoomBlur, Fog, Frost, Snow, Spatter, Contrast, Brightness, Saturate, JpegCompression, Pixelate, ElasticTransform. The augmenters are accessible via iaa.imgcorruptlike.<AugmenterName>.
  • Wrappers around PIL functions (verified to have identical outputs): Solarize, Posterize, Equalize, Autocontrast, EnhanceColor, EnhanceContrast, EnhanceBrightness, EnhanceSharpness, FilterBlur, FilterSmooth, FilterSmoothMore, FilterEdgeEnhance, FilterEdgeEnhanceMore FilterFindEdges, FilterContour, FilterEmboss, FilterSharpen, FilterDetail, Affine. The augmenters are accessible via iaa.pillike.<AugmenterName>.

Aside from these new augmenters, the following major changes were made:

  • Bounding boxes and line strings have now native augmentation methods. Previously, they were converted to keypoints at the start of the augmentation, which meant that child augmenters were unable to use augmentation routines geared towards these two input types as they would merely see a bunch of keypoints.
  • Augmentation happens now batchwise. Previously it was done input type-wise. This change should improve performance for batches with different types of inputs by re-using computation results for multiple inputs. It also makes the library more flexible.
  • Improved the default parameters of augmenters. Most of them will now produce medium-strength augmentations when instantiated without and parameters. E.g. CoarseDropout() will now produce decent augmentations instead of doing nothing. When using default parameters, Fliplr() and Flipud() will always flip (p=100%). TotalDropout() will always drop everything (p=100%). Grayscale() and RemoveSaturation() will always fully grayscale/desaturate. Rot90() will always rotate once (clockwise). Invert() will always invert all components (p=100%).
  • Reworked the standard parameters shared by all augmenters. random_state was renamed to seed, e.g. Affine(..., seed=1) is now valid. The parameter deterministic is now deprecated.
  • Many methods were added to augmentables, e.g. BoundingBoxesOnImage now supports index-based access (bbs[0] instead of bbs.bounding_boxes[0]).
  • The bounding box drawing methods now also draw each BB's label.
  • All augmenters are now tested to be pickle-able without errors.
  • The library is now compatible with numpy 1.18 and python 3.8.
  • This release fixes two significant bugs in Affine that could lead to unaligned outputs. It also fixes significant bugs related to bounding box augmentation and various other issues. The update is recommended. There are now around 5000 unique tests.

Example Images

Brightness

Three new brightness-related augmenters are introduced. The example below shows AddToBrightness. First image is the input, the others show AddToBrightness(-100) to AddToBrightness(100).

Cartoon

A new cartoon style filter is introduced, shown below. Each row starts with the input image.

ChangeColorTemperature

The color temperature of images can now be modified. The example below shows ChangeColorTemperature(kelvin=1000) to ChangeColorTemperature(kelvin=5000), with the first image being the input.

Cutout

Cutout is added to the library. The first row shows the hyperparameters that were used in the corresponding paper. The second row shows two cutout iterations per image, using intensity values, random RGB values and gaussian noise to fill in the pixels. First image in each row is the input.

Dropout2d and TotalDropout

Two new dropout augmenters, Dropout2d and TotalDropout, are added. The example below shows Dropout2d. First image is the input.

Jigsaw

A jigsaw puzzle augmenter is added. The first row below shows its effects using a grid size of 5x5. The second row shows 10x10. First image in each row is the input.

MeanShiftBlur

A mean shift-based blur augmenter is added. First image below shows the input, followed by MeanShiftBlur(5.0) to MeanShiftBlur(40.0).

Posterize

The example below shows Posterize (aka UniformQuantizationToNBits) with n_bits=8 to n_bits=1. First image is the input.

Solarize

The example below shows Solarize, which is the same as Invert with a threshold. First image is the input.

Rain

The example below shows the new Rain augmenter. First image is the input.

RandAugment

This release adds an implementation of RandAugment the following example shows RandAugment(n=2, m=20). First image is the input.

WithPolarWarping

The example below shows WithPolarWarping(<children>) in combination with CropAndPad (first row), Affine (second row) and AveragePooling (third row). First image in each row is the input.

The augmenter supports all input types, but bounding boxes and polygons should be used with caution. (Bounding boxes, because they tend to produce unintuitive results in combination with rotation-like augmentations. Polygons, because they can become invalid under geometric augmentations and will have to be repaired, which can easily mess them up.)

imagecorruptions wrappers

Wrappers around the library imagecorruptions are added, which contains augmentation methods introduced by Hendrycks and Dietterich - Benchmarking Neural Network Robustness to Common Corruptions and Surface Variations. The methods were used in some recent papers. The example below shows their effects, always with severity=3.

PIL wrappers

Various wrappers around popular PIL methods are added. The image below shows in the first row Autocontrast, in the second EnhanceColor (strength of 0.1 to 1.9), the third EnhanceSharpness (strength of 0.1 to 1.9), the fourth shows various convolution-based filters (FilterBlur, FilterSmooth, FilterEdgeEnhance, FilterFindEdges, FilterContour, FilterSharpen, FilterDetail -- in that order) and the fourth row shows pillike.Affine with the top-left as the transformation origin. The first image in each row is the input.

More Blending Augmenters

Various new (alpha-)blending augmenters are introduced in this patch.

The following example makes use of a segmentation map in which all cars are marked with a segmentation class id. It uses roughly BlendAlphaSegMapClassIds(BlendAlphaSomeColors(AddToHueAndSaturation(...))) in order to modify some colors within the car classes. Left is the input image, right is the output:

Note that BlendAlphaSegMapClassIds must be called with all inputs at the same time, e.g. via augmenters(images=..., segmentation_maps...).

This example changes the train color using BlendAlphaSegMapClassIds(AddToHueAndSaturation(...)). The train has a separate class in the segmentation map.

The next example applies blending to some non color-based augmenters. It uses roughly BlendAlphaSegMapClassIds(AdditiveGaussianNoise(...)) (left) and BlendAlphaSegMapClassIds(Emboss(...)) (right). The street has a separate class in the segmentation map.

This example shows how blending can be used to achieve dropout effects. It uses roughly BlendAlphaRegularGrid(Multiply(0.0)) (left) and BlendAlphaCheckerboard(Multiply(0.0)) (right).

This example shows BlendAlphaSomeColors(RemoveSaturation(1.0)), applied to a more colorful image:

This release also adds BlendAlphaBoundingBoxes, BlendAlphaHorizontalLinearGradient and BlendAlphaVerticalLinearGradient. These are not visualized here.

SaveDebugImageEveryNBatches

A new debug helper -- SaveDebugImageEveryNBatches -- was added. The example below shows one of its outputs for a batch containing images, segmentation maps and bounding boxes.

Note that this augmenter must be called with all inputs at the same time, e.g. via augmenters(images=..., segmentation_maps...) for image + segmap inputs.

Mixed-Category Patches

Reworked Augmentation Methods #451 #566

The internal backend of the library was changed so that augmentation now happens batchwise instead of input-type-wise. Child augmenters still have the option of using input-type-wise augmentation. All calls are now at some point routed through Augmenter.augment_batch_() and child augmenters are expected to implement _augment_batch_(). This change allows to re-use information between different input types within the same batch, which in turn improves performance and extends the space of possible augmentations.

Note: It is now recommended to use a batch-wise augmentation call. I.e. use .augment_batch_() or .augment() or .__call__(). These calls provide all inputs of a batch at the same time and several of the new augmenters now explicitly require that (e.g. BlendAlphaBoundingBoxes). Example:

import numpy as np
import imgaug as ia
import imgaug.augmenters as iaa

images = [np.zeros((32, 32, 3), dtype=np.uint8),
          np.zeros((64, 64, 3), dtype=np.uint8)]
bbs = [
    [ia.BoundingBox(x1=0, y1=1, x2=2, y2=3)],
    [ia.BoundingBox(x1=1, y1=2, x2=3, y2=4),
     ia.BoundingBox(x1=2, y1=3, x2=4, y2=5)],
]

bbsois = [ia.BoundingBoxesOnImage(bbs[0], shape=images[0]),
          ia.BoundingBoxesOnImage(bbs[1], shape=images[1])]

aug = iaa.Affine(rotate=(-30, 30))

# No longer recommended:
aug_det = aug.to_deterministic()
images_aug = aug_det.augment_images(images)
bbsois_aug = aug_det.augment_bounding_boxes(bbsois)

# Now recommended:
images_aug, bbs_aug = aug(images=images, bounding_boxes=bbs)

  • Added methods:

    • augmentables.batches.Batch.to_normalized_batch().
    • augmentables.batches.Batch.get_augmentables().
    • augmentables.batches.UnnormalizedBatch.get_augmentables().
    • augmentables.batches.Batch.get_augmentable_names().
    • augmentables.batches.UnnormalizedBatch.get_augmentable_names().
    • augmentables.batches.Batch.to_batch_in_augmentation().
    • augmentables.batches.Batch.fill_from_batch_in_augmentation_().
    • augmentables.batches.UnnormalizedBatch.fill_from_augmented_normalized_batch().
    • augmenters.meta.Augmenter.augment_batch_(), , similar to augment_batch(), but explicitly works in-place and has a parent parameter.
    • augmenters.meta.Augmenter._augment_batch_().
    • augmentables.polys.recover_psois_().
    • augmentables.utils.convert_cbaois_to_kpsois().
    • augmentables.utils.invert_convert_cbaois_to_kpsois_().
    • augmentables.utils.deepcopy_fast().
    • augmentables.bbs.BoundingBox.from_point_soup().
    • augmentables.bbs.BoundingBoxesOnImages.from_point_soups().
    • Added method to_xy_array() to:
      • augmentables.bbs.BoundingBoxesOnImage.
      • augmentables.polys.PolygonsOnImage.
      • augmentables.lines.LineStringsOnImage.
    • Added methods to_keypoints_on_image(), invert_to_keypoints_on_image_() and fill_from_xy_array_() to:
      • augmentables.kps.KeypointsOnImage.
      • augmentables.bbs.BoundingBoxesOnImage.
      • augmentables.polys.PolygonsOnImage.
      • augmentables.lines.LineStringsOnImage.

  • Added classes:

    • testutils.TemporaryDirectory (context)

  • Changed:

    • Changed the following methods to be thin wrappers around augment_batch_():
      • augmenters.meta.Augmenter.augment_images()
      • augmenters.meta.Augmenter.augment_heatmaps().
      • augmenters.meta.Augmenter.augment_segmentation_maps().
      • augmenters.meta.Augmenter.augment_keypoints().
      • augmenters.meta.Augmenter.augment_bounding_boxes().
      • augmenters.meta.Augmenter.augment_polygons().
      • augmenters.meta.Augmenter.augment_line_strings().
    • Changed augment_image(), augment_images(), augment_heatmaps(), augment_segmentation_maps(), augment_keypoints(), augment_bounding_boxes(), augment_polygons() and augment_line_strings() to return None inputs without change. Previously they resulted in an exception. This is more consistent with the behaviour in the other augment_* methods.
    • Changed augment_images() to no longer be abstract. It defaults to not changing the input images.
    • Changed imgaug.augmentables.BoundingBoxesOnImage.from_xyxy_array() to also accept (N, 2, 2) arrays instead of only (N, 4).

Deprecated:

  • Deprecated imgaug.augmenters.meta.Augmenter.augment_batch(). Use .augment_batch_() instead.

Refactored:

  • Refactored most augmenters to use single _augment_batch_() method.

Other changes:

  • Added validation of input arguments to KeypointsOnImage.from_xy_array().
  • Improved validation of input arguments to BoundingBoxesOnImage.from_xyxy_array().

Reworked Quantization #467

This patch reworked the quantization routines to also support quantization to N bits instead of N colors in a way that is similar to posterization in PIL. The patch added corresponding UniformColorQuantizationToNBits and Posterize augmenters, as well as a quantize_uniform_to_n_bits() function.

  • Added classes:

    • augmenters.color.UniformColorQuantizationToNBits.
    • augmenters.color.Posterize (alias of UniformColorQuantizationToNBits).

  • Added functions:

    • augmenters.color.quantize_uniform_(), the in-place version of quantize_uniform().
    • augmenters.color.quantize_uniform_to_n_bits().
    • augmenters.color.quantize_uniform_to_n_bits_().
    • augmenters.color.posterize(), an alias of quantize_uniform_to_n_bits() that produces the same outputs as PIL.ImageOps.posterize().

  • Added parameters:

    • Added to_bin_centers=True to quantize_uniform(), controling whether each bin (a, b) should be quantized to a + (b-a)/2 or a.

  • Deprecated:

    • Renamed imgaug.augmenters.color.quantize_colors_uniform(image, n_colors) to imgaug.augmenters.color.quantize_uniform(arr, nb_bins). The old name is now deprecated.
    • Renamed imgaug.augmenters.color.quantize_colors_kmeans(image, n_colors) to imgaug.augmenters.color.quantize_kmeans(arr, nb_clusters). The old name is now deprecated.

  • Other changes:

    • Improved performance of quantize_uniform() by roughly 10x (small images around 64x64) to 100x (large images around 1024x1024). This also affects UniformColorQuantization.
    • Improved performance of UniformColorQuantization by using more in-place functions.

  • Fixed:

    • Fixed quantize_uniform() producing wrong outputs for non-contiguous arrays.

Improved Invert #469

Added thresholds to Invert and the corresponding functions. This enables solarization (inversion with thresholds). The patch also added a corresponding Solarize augmenter and two solarization functions.

  • Added augmenter imgaug.augmenters.Solarize, a wrapper around Invert.
  • Added function imgaug.augmenters.arithmetic.solarize(), a wrapper around solarize_().
  • Added function imgaug.augmenters.arithmetic.solarize_(), a wrapper around invert_().
  • Added function imgaug.augmenters.arithmetic.invert_(), an in-place version of imgaug.augmenters.arithmetic.invert().
  • Added parameters threshold and invert_above_threshold to imgaug.augmenters.arithmetic.invert()
  • Added parameters threshold and invert_above_threshold to imgaug.augmenters.arithmetic.Invert.
  • Improved performance of imgaug.augmenters.arithmetic.invert() and imgaug.augmenters.arithmetic.Invert for uint8 images.

All Augmenters are now Pickle-able #493 #575

Ensured that all augmenters can be pickled and un-pickled without errors.

  • Added function imgaug.testutils.runtest_pickleable_uint8_img().
  • Fixed imgaug.augmenters.blur.MotionBlur not being pickle-able.
  • Fixed imgaug.augmenters.meta.AssertLambda not being pickle-able.
  • Fixed imgaug.augmenters.meta.AssertShape not being pickle-able.
  • Fixed imgaug.augmenters.color.MultiplyHueAndSaturation not supporting all standard RNG datatypes for random_state.

Extended Cropping and Padding Augmenters #459

This patch extended the cropping and padding augmenters. It added augmenters that crop/pad towards multiples of values (e.g. crop the width until it is a multiple of 2), towards powers of values (e.g. crop the width until it is one of 1, 2, 4, 8, 16, ...), towards an aspect ratio (crop the width or height until width/height = 2.0) or towards a squared size (e.g. crop the width or height until they are equal). These augmenters are wrappers around CropToFixedSize and PadToFixedSize. All *FixedSize augmenters also have now corresponding Center*ToFixedSize aliases, e.g. CenterCropToPowersOf. These are equivalent to using position="center", e.g. CenterCropToPowersOf is equivalent to CropToPowersOf(..., position="center").

The following functions were moved. Their old names are now deprecated.

  • Moved imgaug.imgaug.pad to imgaug.augmenters.size.pad
  • Moved imgaug.imgaug.pad_to_aspect_ratio to imgaug.augmenters.size.pad_to_aspect_ratio.
  • Moved imgaug.imgaug.pad_to_multiples_of to imgaug.augmenters.size.pad_to_multiples_of.
  • Moved imgaug.imgaug.compute_paddings_for_aspect_ratio to imgaug.augmenters.size.compute_paddings_to_reach_aspect_ratio.
  • Moved imgaug.imgaug.compute_paddings_to_reach_multiples_of to imgaug.augmenters.size.compute_paddings_to_reach_multiples_of.

The following augmenters were added:

  • Added augmenter CenterCropToFixedSize.
  • Added augmenter CenterPadToFixedSize.
  • Added augmenter CropToMultiplesOf.
  • Added augmenter CenterCropToMultiplesOf.
  • Added augmenter PadToMultiplesOf.
  • Added augmenter CenterPadToMultiplesOf.
  • Added augmenter CropToPowersOf.
  • Added augmenter CenterCropToPowersOf.
  • Added augmenter PadToPowersOf.
  • Added augmenter CenterPadToPowersOf.
  • Added augmenter CropToAspectRatio.
  • Added augmenter CenterCropToAspectRatio.
  • Added augmenter PadToAspectRatio.
  • Added augmenter CenterPadToAspectRatio.
  • Added augmenter PadToSquare.
  • Added augmenter CenterPadToSquare.
  • Added augmenter CropToSquare.
  • Added augmenter CenterCropToSquare.

All Center<name> augmenters are wrappers around <name> with parameter position="center".

Added functions:

  • Added function imgaug.augmenters.size.compute_croppings_to_reach_aspect_ratio().
  • Added function imgaug.augmenters.size.compute_croppings_to_reach_multiples_of().
  • Added function imgaug.augmenters.size.compute_croppings_to_reach_powers_of().
  • Added function imgaug.augmenters.size.compute_paddings_to_reach_powers_of().

Other changes:

  • Extended augmenter CropToFixedSize to support height and/or width parameters to be None, in which case the respective axis is not changed.
  • Extended augmenter PadToFixedSize to support height and/or width parameters to be None, in which case the respective axis is not changed.
  • [rarely breaking] Changed CropToFixedSize.get_parameters() to also return the height and width values.
  • [rarely breaking] Changed PadToFixedSize.get_parameters() to also return the height and width values.
  • [rarely breaking] Changed the order of parameters returned by PadToFixedSize.get_parameters() to match the order in PadToFixedSize.__init__()
  • Changed PadToFixedSize to prefer padding the right side over the left side and the bottom side over the top side. E.g. if using a center pad and 3 columns have to be padded, it will pad 1 on the left and 2 on the right. Previously it was the other way round. This was changed to establish more consistency with the various other pad and crop methods.
  • Changed the projection of pad/crop values between images and non-images to make the behaviour slightly more accurate in fringe cases.
  • Improved behaviour of function imgaug.augmenters.size.compute_paddings_for_aspect_ratio() for zero-sized axes.
  • Changed function imgaug.augmenters.size.compute_paddings_for_aspect_ratio() to also support shape tuples instead of only ndarrays.
  • Changed function imgaug.augmenters.size.compute_paddings_to_reach_multiples_of() to also support shape tuples instead of only ndarrays.

Fixes:

  • Fixed a formatting error in an error message of compute_paddings_to_reach_multiples_of().

More Choices for Image Blending #462 #556

The available augmenters for alpha-blending of images were significantly extended. There are now new blending augmenters available to alpha-blend acoording to:

  • Some randomly chosen colors. (BlendAlphaSomeColors)
  • Linear gradients. (BlendAlphaHorizontalLinearGradient, BlendAlphaVerticalLinearGradient)
  • Regular grids and checkerboard patterns. (BlendAlphaRegularGrid, BlendAlphaCheckerboard)
  • Only at locations that overlap with specific segmentation class IDs (or the inverse of that). (BlendAlphaSegMapClassIds)
  • Only within bounding boxes with specific labels (or the inverse of that). (BlendAlphaBoundingBoxes)

This allows to e.g. randomly remove some colors while leaving other colors unchanged (BlendAlphaSomeColors(Grayscale(1.0))), to change the color of some objects (BlendAlphaSegMapClassIds(AddToHue((-256, 256)))), to add cloud-patterns only to the top of images (BlendAlphaVerticalLinearGradient(Clouds())) or to apply augmenters in some coarse rectangular areas (e.g. BlendAlphaRegularGrid(Multiply(0.0)) to achieve a similar effect to CoarseDropout or BlendAlphaRegularGrid(AveragePooling(8)) to pool in equally coarse image sub-regions).

Other mask-based alpha blending techniques can be achieved by subclassing IBatchwiseMaskGenerator and providing an instance of such a class to BlendAlphaMask.

This patch also changes the naming of the blending augmenters as follows:

  • Alpha -> BlendAlpha
  • AlphaElementwise -> BlendAlphaElementwise
  • SimplexNoiseAlpha -> BlendAlphaSimplexNoise
  • FrequencyNoiseAlpha -> BlendAlphaFrequencyNoise The old names are now deprecated. Furthermore, the parameters first and second, which were used by all blending augmenters, have now the names foreground and background.

List of changes:

  • Added imgaug.augmenters.blend.BlendAlphaMask, which uses a mask generator instance to generate per batch alpha masks and then alpha-blends using these masks.
  • Added imgaug.augmenters.blend.BlendAlphaSomeColors.
  • Added imgaug.augmenters.blend.BlendAlphaHorizontalLinearGradient.
  • Added imgaug.augmenters.blend.BlendAlphaVerticalLinearGradient.
  • Added imgaug.augmenters.blend.BlendAlphaRegularGrid.
  • Added imgaug.augmenters.blend.BlendAlphaCheckerboard.
  • Added imgaug.augmenters.blend.BlendAlphaSegMapClassIds.
  • Added imgaug.augmenters.blend.BlendAlphaBoundingBoxes.
  • Added imgaug.augmenters.blend.IBatchwiseMaskGenerator, an interface for classes generating masks on a batch-by-batch basis.
  • Added imgaug.augmenters.blend.StochasticParameterMaskGen, a helper to generate masks from StochasticParameter instances.
  • Added imgaug.augmenters.blend.SomeColorsMaskGen, a generator that produces masks marking randomly chosen colors in images.
  • Added imgaug.augmenters.blend.HorizontalLinearGradientMaskGen, a linear gradient mask generator.
  • Added imgaug.augmenters.blend.VerticalLinearGradientMaskGen, a linear gradient mask generator.
  • Added imgaug.augmenters.blend.RegularGridMaskGen, a checkerboard-like mask generator where every grid cell has a random alpha value.
  • Added imgaug.augmenters.blend.CheckerboardMaskGen, a checkerboard-like mask generator where every grid cell has the opposite alpha value of its 4-neighbours.
  • Added imgaug.augmenters.blend.SegMapClassIdsMaskGen, a segmentation map-based mask generator.
  • Added imgaug.augmenters.blend.BoundingBoxesMaskGen, a bounding box-based mask generator.
  • Added imgaug.augmenters.blend.InvertMaskGen, an mask generator that inverts masks produces by child generators.
  • Changed imgaug.parameters.SimplexNoise and imgaug.parameters.FrequencyNoise to also accept (H, W, C) sampling shapes, instead of only (H, W).
  • Refactored AlphaElementwise to be a wrapper around BlendAlphaMask.
  • Renamed Alpha to BlendAlpha. Alpha is now deprecated.
  • Renamed AlphaElementwise to BlendAlphaElementwise. AlphaElementwise is now deprecated.
  • Renamed SimplexNoiseAlpha to BlendAlphaSimplexNoise. SimplexNoiseAlpha is now deprecated.
  • Renamed FrequencyNoiseAlpha to BlendAlphaFrequencyNoise. FrequencyNoiseAlpha is now deprecated.
  • Renamed arguments first and second to foreground and background in BlendAlpha, BlendAlphaElementwise, BlendAlphaSimplexNoise and BlendAlphaFrequencyNoise.
  • Changed imgaug.parameters.handle_categorical_string_param() to allow parameter valid_values to be None.
  • Fixed a wrong error message in imgaug.augmenters.color.change_colorspace_().

Added

Unwrapped Bounding Box Augmentation #446

The bounding box augmentation was previously a wrapper around keypoint augmentation. Bounding Boxes were simply converted to keypoints at the start of the augmentation and then augmented as keypoints by all called augmenters. This was now changed so that all augmenters receive bounding boxes and can then chose how to augment them. This enables augmentations specific to bounding boxes.

  • Added property coords to BoundingBox. The property returns an (N,2) numpy array containing the coordinates of the top-left and bottom-right bounding box corners.
  • Added method BoundingBox.coords_almost_equals(other).
  • Added method BoundingBox.almost_equals(other).
  • Changed method Polygon.almost_equals(other) to no longer verify the datatype. It is assumed now that the input is a Polygon.
  • Added property items to KeypointsOnImage, BoundingBoxesOnImage, PolygonsOnImage, LineStringsOnImage. The property returns the keypoints/BBs/polygons/LineStrings contained by that instance.
  • Added method Polygon.coords_almost_equals(other). Alias for Polygon.exterior_almost_equals(other).
  • Added property Polygon.coords. Alias for Polygon.exterior.
  • Added property Keypoint.coords.
  • Added method Keypoint.coords_almost_equals(other).
  • Added method Keypoint.almost_equals(other).
  • Added method imgaug.testutils.assert_cbaois_equal().
  • Added method imgaug.testutils.shift_cbaoi().
  • Added internal _augment_bounding_boxes() methods to various augmenters. This allows to individually control how bounding boxes are supposed to be augmented. Previously, the bounding box augmentation was a wrapper around keypoint augmentation that did not allow such control.
  • [breaking] Added parameter parents to Augmenter.augment_bounding_boxes(). This breaks if hooks was used as a positional argument in connection with that method.
  • [rarely breaking] Added parameter func_bounding_boxes to Lambda. This breaks if one relied on the order of the augmenter's parameters instead of their names.
  • [rarely breaking] Added parameter func_bounding_boxes to AssertLambda. This breaks if one relied on the order of the augmenter's parameters instead of their names.
  • [rarely breaking] Added parameter check_bounding_boxes to AssertShape. This breaks if one relied on the order of the augmenter's parameters instead of their names.

Unwrapped Line String Augmentation #450

This patch is the same as the bounding box unwrapping above, only applied to line strings.

  • Added internal _augment_line_strings() methods to various augmenters. This allows to individually control how line strings are supposed to be augmented. Previously, the line string augmentation was a wrapper around keypoint augmentation that did not allow such control.
  • [rarely breaking] Added parameter func_line_strings to Lambda. This breaks if one relied on the order of the augmenter's parameters instead of their names.
  • [rarely breaking] Added parameter func_line_strings to AssertLambda. This breaks if one relied on the order of the augmenter's parameters instead of their names.
  • [rarely breaking] Added parameter check_line_strings to AssertShape. This breaks if one relied on the order of the augmenter's parameters instead of their names.

Added fit_output to PerspectiveTransform #452 #456

This patch added fit_output to PerspectiveTransform.

  • [rarely breaking] PerspectiveTransform has now a fit_output parameter, similar to Affine. This change may break code that relied on the order of arguments to __init__.
  • The sampling code of PerspectiveTransform was reworked and should now be faster.

Added ChangeColorTemperature Augmenter #454

This patch added an augmenter and corresponding function to change the color temperature of images. This adds e.g. red, orange or blue tints.

  • Added augmenter imgaug.augmenters.color.ChangeColorTemperature.
  • Added function imgaug.augmenters.color.change_color_temperatures_().
  • Added function imgaug.augmenters.color.change_color_temperature_().

Added Brightness Augmenters #455

This patch added brightness-related augmenters. At the core is WithBrightnessChannels, which converts images to a choice of colorspaces that have brightness-related channels, extracts these channels and applies child augmenters to them. E.g. it might transform to L*a*b* colorspace and extract L, then apply a child augmenter and convert the modified L*a*b* back to RGB.

  • Added augmenter imgaug.augmenters.color.WithBrightnessChannels.
  • Added augmenter imgaug.augmenters.color.MultiplyAndAddToBrightness.
  • Added augmenter imgaug.augmenters.color.MultiplyBrightness.
  • Added augmenter imgaug.augmenters.color.AddToBrightness.
  • Added method imgaug.parameters.handle_categorical_string_param().
  • Changed change_colorspaces_() to accept any iterable of str for argument to_colorspaces, not just list.

Added More Dropout Augmenters #458

This patch added more dropout augmenters. Dropout2d randomly zeros whole channels, while TotalDropout randomly zeros whole images. The latter augmenter can sometimes be used in connection with blending operations. (Note though that in these cases it should not be used with coordinate-based input data, such as bounding boxes, because it removes that data from examples affected by total dropout. That breaks the blending operation, which requires the number of coordinates to be unchanged.)

  • Added a new augmenter Dropout2d, which drops channels in images with a defineable probability p. Dropped channels will be filled with zeros. By default, the augmenter keeps at least one channel in each image unaltered (i.e. not dropped).
  • Added new augmenter TotalDropout, which sets all components to zero for p percent of all images. The augmenter should be used in connection with e.g. blend augmenters.

Added RemoveSaturation #462

  • Added RemoveSaturation, a shortcut for MultiplySaturation((0.0, 1.0)) with outputs similar to Grayscale((0.0, 1.0)).

Added Cartoon Augmenter #463

This patch added a filter to change the style of images to one that looks more cartoon-ish. The filter used classical methods. As such it works well on some images and badly on others. It seems to work better on images that already have rather saturated colors and pronounced edges.

  • Added module imgaug.augmenters.artistic.
  • Added function imgaug.augmenters.artistic.stylize_cartoon(image).
  • Added augmenter imgaug.augmenters.artistic.Cartoon.

Added MeanShiftBlur Augmenter #466

This patch added a mean shift-based blur filter. Note that it is very slow when using the default parameters (high radius).

  • Added function imgaug.augmenters.blur.blur_mean_shift_(image).
  • Added augmenter imgaug.augmenters.blur.MeanShiftBlur.

Added DeterministicList Parameter #475

Added imgaug.parameters.DeterministicList. Upon a request to generate samples of shape S, this parameter will create a new array of shape S and fill it by cycling over its list of values repeatedly.

Added Jigsaw Augmenter #476 #577

This patch added a jigsaw puzzle augmenter and corresponding functions. The augmenter splits each image into a regular grid of cells, then randomly picks some cells and switches them with one of their 8-neighbours. The process is repeated for N steps per image.

Note: The augmenter will reject batches containing bounding boxes, polygons or line strings.

  • Added function imgaug.augmenters.geometric.apply_jigsaw().
  • Added function imgaug.augmenters.geometric.apply_jigsaw_to_coords().
  • Added function imgaug.augmenters.geometric.generate_jigsaw_destinations().

Added Wrappers around Package PIL #479 #480 #538

This patch added wrapper functions and augmenters around popular PIL functions. The outputs of these functions and augmenters are tested to be identical with the ones in PIL. They are intended for research cases where papers have to be re-implemented as accurately as possible.

  • Added module imgaug.augmenters.pillike, which contains augmenters and functions corresponding to commonly used PIL functions. Their outputs are guaranteed to be identical to the PIL outputs.
  • Added the following functions to the module:
    • imgaug.augmenters.pillike.equalize
    • imgaug.augmenters.pillike.equalize_
    • imgaug.augmenters.pillike.autocontrast
    • imgaug.augmenters.pillike.autocontrast_
    • imgaug.augmenters.pillike.solarize
    • imgaug.augmenters.pillike.solarize_
    • imgaug.augmenters.pillike.posterize
    • imgaug.augmenters.pillike.posterize_
    • imgaug.augmenters.pillike.enhance_color
    • imgaug.augmenters.pillike.enhance_contrast
    • imgaug.augmenters.pillike.enhance_brightness
    • imgaug.augmenters.pillike.enhance_sharpness
    • imgaug.augmenters.pillike.filter_blur
    • imgaug.augmenters.pillike.filter_smooth
    • imgaug.augmenters.pillike.filter_smooth_more
    • imgaug.augmenters.pillike.filter_edge_enhance
    • imgaug.augmenters.pillike.filter_edge_enhance_more
    • imgaug.augmenters.pillike.filter_find_edges
    • imgaug.augmenters.pillike.filter_contour
    • imgaug.augmenters.pillike.filter_emboss
    • imgaug.augmenters.pillike.filter_sharpen
    • imgaug.augmenters.pillike.filter_detail
    • imgaug.augmenters.pillike.warp_affine
  • Added the following augmenters to the module:
    • imgaug.augmenters.pillike.Solarize
    • imgaug.augmenters.pillike.Posterize. (Currently alias for imgaug.augmenters.color.Posterize.)
    • imgaug.augmenters.pillike.Equalize
    • imgaug.augmenters.pillike.Autocontrast
    • imgaug.augmenters.pillike.EnhanceColor
    • imgaug.augmenters.pillike.EnhanceContrast
    • imgaug.augmenters.pillike.EnhanceBrightness
    • imgaug.augmenters.pillike.EnhanceSharpness
    • imgaug.augmenters.pillike.FilterBlur
    • imgaug.augmenters.pillike.FilterSmooth
    • imgaug.augmenters.pillike.FilterSmoothMore
    • imgaug.augmenters.pillike.FilterEdgeEnhance
    • imgaug.augmenters.pillike.FilterEdgeEnhanceMore
    • imgaug.augmenters.pillike.FilterFindEdges
    • imgaug.augmenters.pillike.FilterContour
    • imgaug.augmenters.pillike.FilterEmboss
    • imgaug.augmenters.pillike.FilterSharpen
    • imgaug.augmenters.pillike.FilterDetail
    • imgaug.augmenters.pillike.Affine

Added Identity #481

This patch added an identity function augmenter (Identity), which is the same as Noop and will replace the latter one in the long run.

  • [rarely breaking] Added imgaug.augmenters.meta.Identity, an alias of Noop. Identity is now the recommended augmenter for identity transformations. This change can break code that explicitly relied on exactly Noop being used, e.g. via isinstance checks.
  • Renamed parameter noop_if_topmost to identity_if_topmost in method imgaug.augmenters.meta.Augmenter.remove_augmenters(). The old name is now deprecated.

Added Shearing on the Y-Axis to Affine #482

Affine was changed to now also support shearing on the y-axis. Previously, only the x-axis was supported. Use e.g. Affine(shear={"y": (-20, 20)) now.

  • [rarely breaking] Extended Affine to also support shearing on the y-axis (previously, only x-axis was possible). This feature can be used via e.g. Affine(shear={"x": (-30, 30), "y": (-10, 10)}). If instead a single number is used (e.g. Affine(shear=15)), shearing will be done only on the x-axis. If a single tuple, list or StochasticParameter is used, the generated samples will be used identically for both the x-axis and y-axis (this is consistent with translation and scaling). To get independent random samples per axis use the dictionary form.

Added Wrappers around Affine #484

This patch added a few convenience wrappers around Affine.

  • Added imgaug.augmenters.geometric.ScaleX.
  • Added imgaug.augmenters.geometric.ScaleY.
  • Added imgaug.augmenters.geometric.TranslateX.
  • Added imgaug.augmenters.geometric.TranslateY.
  • Added imgaug.augmenters.geometric.Rotate.
  • Added imgaug.augmenters.geometric.ShearX.
  • Added imgaug.augmenters.geometric.ShearY.

Added More Methods to Remove Out-of-Image Augmentables #487

This patch extended the methods to handle coordinate-based augmentables, e.g. bounding boxes, that are partially/fully outside of the image plane. They can now more easily be dropped if more than p% of their areas is outside of the image plane.

The patch also adds augmenters to remove and clip coordinate-based augmentables that are outside of the image plane.

  • Added Keypoint.is_out_of_image().

  • Added BoundingBox.compute_out_of_image_area().

  • Added Polygon.compute_out_of_image_area().

  • Added Keypoint.compute_out_of_image_fraction()

  • Added BoundingBox.compute_out_of_image_fraction().

  • Added Polygon.compute_out_of_image_fraction().

  • Added LineString.compute_out_of_image_fraction().

  • Added KeypointsOnImage.remove_out_of_image_fraction().

  • Added BoundingBoxesOnImage.remove_out_of_image_fraction().

  • Added PolygonsOnImage.remove_out_of_image_fraction().

  • Added LineStringsOnImage.remove_out_of_image_fraction().

  • Added KeypointsOnImage.clip_out_of_image().

  • Added imgaug.augmenters.meta.RemoveCBAsByOutOfImageFraction. Removes coordinate-based augmentables (e.g. BBs) that have at least a specified fraction of their area outside of the image plane.

  • Added imgaug.augmenters.meta.ClipCBAsToImagePlanes. Clips off all parts from coordinate-based augmentables (e.g. BBs) that are outside of the corresponding image.

  • Changed Polygon.area to return 0.0 if the polygon contains less than three points (previously: exception).

Added Bounding Box to Polygon Conversion #489

  • Added method imgaug.augmentables.bbs.BoundingBox.to_polygon().
  • Added method imgaug.augmentables.bbs.BoundingBoxesOnImage.to_polygons_on_image().

Added Polygon Subdivision #489

  • Added method imgaug.augmentables.polys.Polygon.subdivide(N). The method increases the polygon's corner point count by interpolating N points on each edge with regular distance.
  • Added method imgaug.augmentables.polys.PolygonsOnImage.subdivide(N).

Added WithPolarWarping Augmenter #489

This patch added an augmenter to transform images to polar coordinates and apply child augmenters within that space. This leads to interesting effects in combination with augmenters that affect pixel locations, such as cropping or affine transformations.

  • Added augmenter imgaug.augmenters.geometric.WithPolarWarping, an augmenter that applies child augmenters in a polar representation of the image.

Added Convenient Access Methods to Coordinate-Based Augmentables #495 #541

This patch added various magic functions to coordinate-based augmentables that make their usage more convenient. Example:

import imgaug as ia
bb1 = ia.BoundingBox(x1=0, y1=1, x2=2, y2=3)
bb2 = ia.BoundingBox(x1=1, y1=2, x2=3, y2=4)
bbsoi = ia.BoundingBoxesOnImage([bb1, bb2])

print(bbsoi[0])  # prints now str(bb1)
print(len(bbsoi))  # prints now 2
for bb in bbsoi:  # looping is now supported 
    print(bb)
  • Added module imgaug.augmentables.base.
  • Added interface imgaug.augmentables.base.IAugmentable, implemented by HeatmapsOnImage, SegmentationMapsOnImage, KeypointsOnImage, BoundingBoxesOnImage, PolygonsOnImage and LineStringsOnImage.
  • Added ability to iterate over coordinate-based *OnImage instances (keypoints, bounding boxes, polygons, line strings), e.g. bbsoi = BoundingBoxesOnImage(bbs, shape=...); for bb in bbsoi: .... would iterate now over bbs.
  • Added implementations of __len__ methods to coordinate-based *OnImage instances, e.g. bbsoi = BoundingBoxesOnImage(bbs, shape=...); print(len(bbsoi)) would now print the number of bounding boxes in bbsoi.
  • Added ability to iterate over coordinates of BoundingBox (top-left, bottom-right), Polygon and LineString via for xy in obj: ....
  • Added ability to access coordinates of BoundingBox, Polygon and LineString using indices or slices, e.g. line_string[1:] to get an array of all coordinates except the first one.
  • Added property Keypoint.xy.
  • Added property Keypoint.xy_int.

Added SaveDebugImageEveryNBatches Augmenter #502

This patch added a debug augmenter SaveDebugImageEveryNBatches that visualizes a whole batch and saves the corresponding image to a directory. The visualization happens at every Nth batch. The plot contains a visualization of all images within the batch, as well as all additional input data (e.g. segmentation maps or bounding boxes overlayed with images). The plot also contains various additional information, such as observed value ranges (min/max values) of images, observed labels of bounding boxes or observed segmentation classes.

The augmenter can be used during training to evaluate the strength of augmentations, whether all data is still aligned (e.g. bounding box positions match object positions) and whether the data statistics match the expectations (e.g. no segmentation map classes missing).

The augmenter might be useful even if no augmentation is actually performed.

  • Added module imgaug.augmenters.debug.
  • Added function imgaug.augmenters.debug.draw_debug_image(). The function draws an image containing debugging information for a provided set of images and non-image data (e.g. segmentation maps, bounding boxes) corresponding to a single batch. The debug image visualizes these informations (e.g. bounding boxes drawn on images) and offers relevant information (e.g. actual value ranges of images, labels of bounding boxes and their counts, etc.).
  • Added augmenter imgaug.augmenters.debug.SaveDebugImageEveryNBatches. Augmenter corresponding to draw_debug_image(). Saves an image at every n-th batch into a provided folder.

Added Multi-Channel cvals in pad() #502

Improved imgaug.augmenters.size.pad() to support multi-channel values for the cval parameter (e.g. RGB colors).

Added Wrappers around Package imagecorruptions #530

Added wrappers around the functions from package bethgelab/imagecorruptions. The functions in that package were used in some recent papers and are added here for convenience. The wrappers produce arrays containing values identical to the output arrays from the corresponding imagecorruptions functions when called via the imagecorruptions.corrupt() (verified via unittests). The interfaces of the wrapper functions are identical to the imagecorruptions functions, with the only difference of also supporting seed parameters.

  • Added module imgaug.augmenters.imgcorruptlike. The like signals that the augmentation functions do not have to wrap imagecorruptions internally. They merely have to produce the same outputs.
  • Added the following functions to module imgaug.augmenters.imgcorruptlike:
    • apply_gaussian_noise()
    • apply_shot_noise()
    • apply_impulse_noise()
    • apply_speckle_noise()
    • apply_gaussian_blur()
    • apply_glass_blur() (improved performance over original function)
    • apply_defocus_blur()
    • apply_motion_blur()
    • apply_zoom_blur()
    • apply_fog()
    • apply_snow()
    • apply_spatter()
    • apply_contrast()
    • apply_brightness()
    • apply_saturate()
    • apply_jpeg_compression()
    • apply_pixelate()
    • apply_elastic_transform()
  • Added function imgaug.augmenters.imgcorruptlike.get_corruption_names(subset). Similar to imagecorruptions.get_corruption_names(subset), but returns a tuple (list of corruption method names, list of corruption method functions), instead of only the names.
  • Added the following augmenters to module imgaug.augmenters.imgcorruptlike:
    • GaussianNoise
    • ShotNoise
    • ImpulseNoise
    • SpeckleNoise
    • GaussianBlur
    • GlassBlur
    • DefocusBlur
    • MotionBlur
    • ZoomBlur
    • Fog
    • Frost
    • Snow
    • Spatter
    • Contrast
    • Brightness
    • Saturate
    • JpegCompression
    • Pixelate
    • ElasticTransform
  • Added context imgaug.random.temporary_numpy_seed().

Added Cutout Augmenter #531 #570

This patch added Cutout augmentation, similar to the paper proposal. The augmetner has some similarity with CoarseDropout.

  • Added imgaug.augmenters.arithmetic.apply_cutout_(), which replaces in-place a single rectangular area with a constant intensity value or a constant color or gaussian noise. See also the paper about Cutout.
  • Added imgaug.augmenters.arithmetic.apply_cutout(). Same as apply_cutout_(), but copies the input images before applying cutout.
  • Added imgaug.augmenters.arithmetic.Cutout.

Added in-place Methods for Coordinate-based Augmentables #532

This patch added for many already existing methods corresponding in-place variations. They are now used throughout the library, improving the performance of augmentation in the case of e.g. bounding boxes.

  • Added Keypoint.project_().
  • Added Keypoint.shift_().
  • Added KeypointsOnImage.on_().
  • Added setter for KeypontsOnImage.items.
  • Added setter for BoundingBoxesOnImage.items.
  • Added setter for LineStringsOnImage.items.
  • Added setter for PolygonsOnImage.items.
  • Added KeypointsOnImage.remove_out_of_image_fraction_().
  • Added KeypointsOnImage.clip_out_of_image_fraction_().
  • Added KeypointsOnImage.shift_().
  • Added BoundingBox.project_().
  • Added BoundingBox.extend_().
  • Added BoundingBox.clip_out_of_image_().
  • Added BoundingBox.shift_().
  • Added BoundingBoxesOnImage.on_().
  • Added BoundingBoxesOnImage.clip_out_of_image_().
  • Added BoundingBoxesOnImage.remove_out_of_image_().
  • Added BoundingBoxesOnImage.remove_out_of_image_fraction_().
  • Added BoundingBoxesOnImage.shift_().
  • Added imgaug.augmentables.utils.project_coords_().
  • Added LineString.project_().
  • Added LineString.shift_().
  • Added LineStringsOnImage.on_().
  • Added LineStringsOnImage.remove_out_of_image_().
  • Added LineStringsOnImage.remove_out_of_image_fraction_().
  • Added LineStringsOnImage.clip_out_of_image_().
  • Added LineStringsOnImage.shift_().
  • Added Polygon.project_().
  • Added Polygon.shift_().
  • Added Polygon.on_().
  • Added Polygon.subdivide_().
  • Added PolygonsOnImage.remove_out_of_image_().
  • Added PolygonsOnImage.remove_out_of_image_fraction_().
  • Added PolygonsOnImage.clip_out_of_image_().
  • Added PolygonsOnImage.shift_().
  • Added PolygonsOnImage.subdivide_().
  • Switched BoundingBoxesOnImage.copy() to a custom copy operation (away from module copy module).
  • Added parameters bounding_boxes and shape to BoundingBoxesOnImage.copy()`.
  • Added parameters bounding_boxes and shape to BoundingBoxesOnImage.deepcopy()`.
  • Switched KeypointsOnImage.copy() to a custom copy operation (away from module copy module).
  • Switched PolygonsOnImage.copy() to a custom copy operation (away from module copy module).
  • Added parameters polygons and shape to PolygonsOnImage.copy()`.
  • Added parameters polygons and shape to PolygonsOnImage.deepcopy()`.
  • Switched augmenters to use in-place functions for keypoints, bounding boxes, line strings and polygons.

Added Standardized LUT Methods #542

This patch standardized the handling of lookup tables throughout the library.

  • Added imgaug.imgaug.apply_lut(), which applies a lookup table to an image.
  • Added imgaug.imgaug.apply_lut_(). In-place version of apply_lut().
  • Refactored all augmenters to use these new LUT functions. This likely fixed some so-far undiscovered bugs in augmenters using LUT tables.

Added Drawing of Bounding Box Labels #545

When drawing bounding boxes on images via BoundingBox.draw_on_image() or BoundingBoxesOnImage.draw_on_image(), a box containing the label will now be drawn over each bounding box's rectangle. If the bounding box's label is set to None, the label box will not be drawn. For more detailed control, use BoundingBox.draw_label_on_image().

  • Added method imgaug.augmentables.BoundingBox.draw_label_on_image().
  • Added method imgaug.augmentables.BoundingBox.draw_box_on_image().
  • Changed method imgaug.augmentables.BoundingBox.draw_on_image() to automatically draw a bounding box's label.

Added Index-based Access to Coordinate-based *OnImage Instances #547

Enabled index-based access to coordinate-based *OnImage instances, i.e. to KeypointsOnImage, BoundingBoxesOnImage, LineStringsOnImage and PolygonsOnImage. This allows to do things like bbsoi = BoundingBoxesOnImage(...); bbs = bbsoi[0:2];.

  • Added imgaug.augmentables.kps.KeypointsOnImage.__getitem__().
  • Added imgaug.augmentables.bbs.BoundingBoxesOnImage.__getitem__().
  • Added imgaug.augmentables.lines.LineStringsOnImage.__getitem__().
  • Added imgaug.augmentables.polys.PolygonsOnImage.__getitem__().

Added Rain and RainLayer Augmenters #551

Added augmenter(s) to create fake rain effects. They currently seem to work best at around medium-sized images (~224px).

  • Added imgaug.augmenters.weather.Rain.
  • Added imgaug.augmenters.weather.RainLayer.

Added round Parameter to Discretize #553

Added the parameter round to imgaug.parameters.Discretize. The parameter defaults to True, i.e. the default behaviour of Discretize did not change.

Added RandAugment Augmenter #553

Added a RandAugment augmenter, similar to the one described in the paper "RandAugment: Practical automated data augmentation with a reduced search space".

Note: This implementation makes a best guess about some hyperparameters that were neither in the paper nor in the code repsitory clearly defined.

Note: This augmenter differs from the paper implementation by applying a fix to their color augmentations. The ones in the paper's implementation seemed to increase in strength as the magnitude was decreased below a threshold.

Note: This augmenter currently only accepts image inputs. Other input types (e.g. bounding boxes) will be rejected.

  • Added module imgaug.augmenters.collections
  • Added augmenter imgaug.augmenters.collections.RandAugment.

Added and Improved Warnings for Probably-Wrong Image Inputs #594

Improved the errors and warnings on image augmentation calls. augment_image() will now produce a more self-explanatory error message when calling it as in augment_image(list of images). Calls of single-image augmentation functions (e.g. augment(image=...)) with inputs that look like multiple images will now produce warnings. This is the case for (H, W, C) inputs when C>=32 (as that indicates that (N, H, W) was actually provided). Calls of multi-image augmentation functions (e.g. augment(images=...)) with inputs that look like single images will now produce warnings. This is the case for (N, H, W) inputs when W=1 or W=3 (as that indicates that (H, W, C) was actually provided.)

  • Added an assert in augment_image() to verify that inputs are arrays.
  • Added warnings for probably-wrong image inputs in augment_image(), augment_images(), augment() (and its alias __call__()).
  • Added module imgaug.augmenters.base.
  • Added warning imgaug.augmenters.base.SuspiciousMultiImageShapeWarning.
  • Added warning imgaug.augmenters.base.SuspiciousSingleImageShapeWarning.
  • Added imgaug.testutils.assertWarns, similar to unittest's assertWarns, but available in python <3.2.

Changed

Improved RNG Handling during Polygon Augmentation #447

Changed Augmenter.augment_polygons() to copy the augmenter's RNG before starting concave polygon recovery. This is done for cleanliness and should not have any effects for users. Also removed RNG copies in _ConcavePolygonRecoverer to improve performance.

Pooling Augmenters now Affect Maps #457

Pooling augmenters were previously implemented so that they did not pool the arrays of maps (i.e. heatmap arrays, segmentation map arrays). Only the image shape saved within HeatmapsOnImage.shape and SegmentationMapsOnImage.shape were updated. That was done because the library can handle map arrays that are larger than the corresponding images and hence no pooling was necessary for the augmentation to work correctly. This was now changed and pooling augmenters will also pool map arrays (if keep_size=False). The motiviation for this change is that the old behaviour was unintuitive and inconsistent with other augmenters (e.g. Crop).

Affine Translation Precision #489

Removed a rounding operation in Affine translation that would unnecessarily round floats to integers. This should make coordinate augmentation overall more accurate.

Affine.get_parameters() and translate_px/translate_percent #508

Changed Affine.get_parameters() to always return a tuple (x, y, mode) for translation, where mode is either px or percent, and x and y are stochastic parameters. y may be None if the same parameter (and hence samples) are used for both axes.

Removed Outdated "Don't Import from this Module" Messages #539

The docstring of each module in imgaug.augmenters previously included a suggestion to not directly import from that module, but instead use imgaug.augmenters.<AugmenterName>. That was due to the categorization still being unstable. As the categorization has now been fairly stable for a long time, the suggestion was removed from all modules. Calling imgaug.augmenters.<AugmenterName> instead of imgaug.augmenters.<ModuleName>.<AugmenterName> is however still the preferred way.

Standardized shift() Interfaces of Coordinate-Based Augmentables #548

The interfaces for shift operations of all coordinate-based augmentables (Keypoints, BoundingBoxes, LineStrings, Polygons) were standardized. All of these augmentables have now the same interface for shift operations. Previously, Keypoints used a different interface (using x and y arguments) than the other augmentables (using top, right, bottom, left arguments). All augmentables use now the interface of Keypoints as that is simpler and less ambiguous. Old arguments are still accepted, but will produce deprecation warnings. Change the arguments to x and y following x=left-right and y=top-bottom.

[breaking] This breaks if one relied on calling shift() functions of BoundingBox, LineString, Polygon, BoundingBoxesOnImage, LineStringsOnImage or PolygonsOnImage without named arguments. E.g. bb = BoundingBox(...); bb_shifted = bb.shift(1, 2, 3, 4); will produce unexpected outputs now (equivalent to shift(x=1, y=2, top=3, right=4, bottom=0, left=0)), while bb_shifted = bb.shift(top=1, right=2, bottom=3, left=4) will still work as expected.

  • Added arguments x, y to BoundingBox.shift(), LineString.shift() and Polygon.shift().
  • Added arguments x, y to BoundingBoxesOnImage.shift(), LineStringsOnImage.shift() and PolygonsOnImage.shift().
  • Marked arguments top, right, bottom, left in BoundingBox.shift(), LineString.shift() and Polygon.shift() as deprecated. This also affects the corresponding *OnImage classes.
  • Added function testutils.wrap_shift_deprecation().

Simplified Standard Parameters of Augmenters #567 #595

The patch changed the standard parameters shared by all augmenters to a reduced and more self-explanatory set. Previously, all augmenters shared the parameters name, random_state and deterministic. The new parameters are seed and name.

deterministic was removed as it was hardly ever used and because it caused frequently confusion with regards to its meaning. The parameter is still accepted but will now produce a deprecation warning. Use <augmenter>.to_deterministic() instead.

Reminder: to_deterministic() is necessary if you want to get the same samples in consecutive augmentation calls. It is not necessary if you want your generated samples to be dependent on an initial seed or random state as that is always the case anyways. To use non-random initial seeds, use either the seed parameter (augmenter-specific seeding) or imgaug.random.seed() (global seeding, affects only augmenters for which the seed parameter was not explicitly provided).

random_state was renamed to seed as providing a seed value is the more common use case compared to providing a random state. Many users also seemed to be unaware that random_state accepted seed values. The new name should make this more clear. The old parameter random_state is still accepted, but will likely be deprecated in the future.

[breaking] This patch breaks if one relied on the order of name, random_state and deterministic. The new order is now seed=..., name=..., random_state=..., deterministic=... (with the latter two parameters being outdated or deprecated) as opposed to previously name=..., deterministic=..., random_state=....

Improved Default Values of Augmenters #582

[breaking] Most augmenters had previously default values that made them equivalent to identity functions. Users had to explicitly change the defaults to proper values in order to "activate" augmentations. To simplify the usage of the library, the default values of most augmenters were changed to medium-strength augmentations. E.g. Sequential([Affine(), UniformVoronoi(), CoarseDropout()]) should now produce decent augmentations.

A few augmenters were set to always-on, maximum-strength augmentations. This is the case for:

  • Grayscale (always fully grayscales images, use Grayscale((0.0, 1.0)) for random strengths)
  • RemoveSaturation (same as Grayscale)
  • Fliplr (always flips images, use Fliplr(0.5) for 50% probability)
  • Flipud (same as Fliplr)
  • TotalDropout (always drops everything, use TotalDropout(0.1) to drop everything for 10% of all images)
  • Invert (always inverts images, use Invert(0.1) to invert 10% of all images)
  • Rot90 (always rotates exactly once clockwise by 90 degrees, use Rot90((0, 3)) for any rotation)

These settings seemed to better match user-expectations. Such maximum-strength settings however were not chosen for all augmenters where one might expect them. The defaults are set to varying strengths for, e.g. Superpixels (replaces only some superpixels with cellwise average colors), UniformVoronoi (also only replaces some cells), Sharpen (alpha-blends with variable strength, the same is the case for Emboss, EdgeDetect and DirectedEdgeDetect) and CLAHE (variable clip limits).

Note: Some of the new default values will cause issues with non-uint8 inputs.

Note: The defaults for per_channel and keep_size were not adjusted. It is currently still the default behaviour of all augmenters to affect all channels in the same way and to resize their outputs back to the input sizes.

The exact changes to default values are listed below.

imgaug.arithmetic

  • Add
    • value: 0 -> (-20, 20)
  • AddElementwise
    • value: 0 -> (-20, 20)
  • AdditiveGaussianNoise
    • scale: 0 -> (0, 15)
  • AdditiveLaplaceNoise
    • scale: 0 -> (0, 15)
  • AdditivePoissonNoise
    • scale: 0 -> (0, 15)
  • Multiply
    • mul: 1.0 -> (0.8, 1.2)
  • MultiplyElementwise:
    • mul: 1.0 -> (0.8, 1.2)
  • Dropout:
    • p: 0.0 -> (0.0, 0.05)
  • CoarseDropout:
    • p: 0.0 -> (0.02, 0.1)
    • size_px: None -> (3, 8)
    • min_size: 4 -> 3
    • Default for size_px is only used if neither size_percent nor size_px is provided by the user.
  • CoarseSaltAndPepper:
    • p: 0.0 -> (0.02, 0.1)
    • size_px: None -> (3, 8)
    • min_size: 4 -> 3
    • Default for size_px is only used if neither size_percent nor size_px is provided by the user.
  • CoarseSalt:
    • p: 0.0 -> (0.02, 0.1)
    • size_px: None -> (3, 8)
    • min_size: 4 -> 3
    • Default for size_px is only used if neither size_percent nor size_px is provided by the user.
  • CoarsePepper:
    • p: 0.0 -> (0.02, 0.1)
    • size_px: None -> (3, 8)
    • min_size: 4 -> 3
    • Default for size_px is only used if neither size_percent nor size_px is provided by the user.
  • SaltAndPepper:
    • p: 0.0 -> (0.0, 0.03)
  • Salt:
    • p: 0.0 -> (0.0, 0.03)
  • Pepper:
    • p: 0.0 -> (0.0, 0.05)
  • ImpulseNoise:
    • p: 0.0 -> (0.0, 0.03)
  • Invert:
    • p: 0 -> 1
  • JpegCompression:
    • compression: 50 -> (0, 100)

imgaug.blend

  • BlendAlpha
    • factor: 0 -> (0.0, 1.0)
  • BlendAlphaElementwise
    • factor: 0 -> (0.0, 1.0)

imgaug.blur

  • GaussianBlur:
    • sigma: 0 -> (0.0, 3.0)
  • AverageBlur:
    • k: 1 -> (1, 7)
  • MedianBlur:
    • k: 1 -> (1, 7)
  • BilateralBlur:
    • d: 1 -> (1, 9)
  • MotionBlur:
    • k: 5 -> (3, 7)

imgaug.color

  • MultiplyHueAndSaturation:
    • mul_hue: None -> (0.5, 1.5)
    • mul_saturation: None -> (0.0, 1.7)
    • These defaults are only used if the user provided neither mul nor mul_hue nor mul_saturation.
  • MultiplyHue:
    • mul: (-1.0, 1.0) -> (-3.0, 3.0)
  • AddToHueAndSaturation:
    • value_hue: None -> (-40, 40)
    • value_saturation: None -> (-40, 40)
    • These defaults are only used if the user provided neither value nor value_hue nor value_saturation.
  • Grayscale:
    • alpha: 0 -> 1

imgaug.contrast

  • GammaContrast:
    • gamma: 1 -> (0.7, 1.7)
  • SigmoidContrast:
    • gain: 10 -> (5, 6)
    • cutoff: 0.5 -> (0.3, 0.6)
  • LogContrast:
    • gain: 1 -> (0.4, 1.6)
  • LinearContrast:
    • alpha: 1 -> (0.6, 1.4)
  • AllChannelsCLAHE:
    • clip_limit: 40 -> (0.1, 8)
    • tile_grid_size_px: 8 -> (3, 12)
  • CLAHE:
    • clip_limit: 40 -> (0.1, 8)
    • tile_grid_size_px: 8 -> (3, 12)

convolutional

  • Sharpen:
    • alpha: 0 -> (0.0, 0.2)
    • lightness: 1 -> (0.8, 1.2)
  • Emboss:
    • alpha: 0 -> (0.0, 1.0)
    • strength: 1 -> (0.25, 1.0)
  • EdgeDetect:
    • alpha: 0 -> (0.0, 0.75)
  • DirectedEdgeDetect:
    • alpha: 0 -> (0.0, 0.75)

imgaug.flip

  • Fliplr:
    • p: 0 -> 1
  • Flipud:
    • p: 0 -> 1

imgaug.geometric

  • Affine:
    • scale: 1 -> {"x": (0.9, 1.1), "y": (0.9, 1.1)}
    • translate_percent: None -> {"x": (-0.1, 0.1), "y": (-0.1, 0.1)}
    • rotate: 0 -> (-15, 15)
    • shear: 0 -> shear={"x": (-10, 10), "y": (-10, 10)}
    • These defaults are only used if no affine transformation parameter was set by the user. Otherwise the not-set parameters default again towards the identity function.
  • PiecewiseAffine:
    • scale: 0 -> (0.0, 0.04)
    • nb_rows: 4 -> (2, 4)
    • nb_cols: 4 -> (2, 4)
  • PerspectiveTransform:
    • scale: 0 -> (0.0, 0.06)
  • ElasticTransformation:
    • alpha: 0 -> (0.0, 40.0)
    • sigma: 0 -> (4.0, 8.0)
  • Rot90:
    • k: (no default) -> k=1

imgaug.pooling

  • AveragePooling:
    • k: (no default) -> (1, 5)
  • MaxPooling:
    • k: (no default) -> (1, 5)
  • MinPooling:
    • k: (no default) -> (1, 5)
  • MedianPooling:
    • k: (no default) -> (1, 5)

imgaug.segmentation

  • Superpixels:
    • p_replace: 0.0 -> (0.5, 1.0)
    • n_segments: 100 -> (50, 120)
  • UniformVoronoi:
    • n_points: (no default) -> (50, 500)
    • p_replace: 1.0 -> (0.5, 1.0).
  • RegularGridVoronoi:
    • n_rows: (no default) -> (10, 30)
    • n_cols: (no default) -> (10, 30)
    • p_drop_points: 0.4 -> (0.0, 0.5)
    • p_replace: 1.0 -> (0.5, 1.0)
  • RelativeRegularGridVoronoi: Changed defaults from
    • n_rows_frac: (no default) -> (0.05, 0.15)
    • n_cols_frac: (no default) -> (0.05, 0.15)
    • p_drop_points: 0.4 -> (0.0, 0.5)
    • p_replace: 1.0 -> (0.5, 1.0)

imgaug.size

  • CropAndPad:
    • percent: None -> (-0.1, 0.1)
    • This default is only used if the user has provided neither px nor percent.
  • Pad:
    • percent: None -> (0.0, 0.1)
    • This default is only used if the user has provided neither px nor percent.
  • Crop:
    • percent: None -> (0.0, 0.1)
    • This default is only used if the user has provided neither px nor percent.

setup.py Now Accepts any opencv-* Installation #586

setup.py was changed so that it now accepts opencv-python, opencv-python-headless, opencv-contrib-python and opencv-contrib-python-headless as valid OpenCV installations. Previously, only opencv-python-headless was accepted, which could easily cause conflicts when another one of the mentioned libraries was already installed. If none of the mentioned libraries is installed, setup.py will default to adding opencv-python as a requirement.

Note that this may still cause issues if a single installation call installs multiple libraries and the order is random. imgaug will then currently request opencv-python-headless to be installed, which may differ from what a later installed library requests. Try to ensure that the other library is installed first in these cases.

Unified OpenCV Input Normalization #565

Changed various augmenters to use the same normalization for OpenCV inputs. This probably fixes some previously undiscovered bugs.

Renamed In-place Methods #444

  • Renamed Augmenter.reseed() to Augmenter.seed_(). The old name is now deprecated.
  • Renamed Augmenter.remove_augmenters_inplace() to Augmenter.remove_augmenters_(). The old name is now deprecated.

Deprecated AffineCv2 #540

Deprecated imgaug.augmenters.geometric.AffineCv2. Use imgaug.augmenters.geometric.Affine instead. #540

Refactored

Refactored According to pylint Requirements #504

  • Refactored all core library files to fulfill (most) pylint requirements.
  • [rarely breaking] Renamed imgaug.augmenters.size.KeepSizeByResize.get_shapes() to _get_shapes().
  • Added a project-specific pylint configuration.

Fixed

  • Fixed Resize always returning an uint8 array during image augmentation if the input was a single numpy array and all augmented images had the same shape. #442 #443
  • Fixed Affine coordinate-based augmentation applying wrong offset when shifting images to/from top-left corner. This would lead to an error of around 0.5 to 1.0 pixels. #446
  • Fixed keypoint augmentation in PiecewiseAffine potentially being unaligned if a KeypointsOnImage instance contained no keypoints. #446
  • Fixed imgaug.validation.convert_iterable_to_string_of_types() crashing due to not converting types to strings before joining them. #446
  • Fixed imgaug.validation.assert_is_iterable_of() producing a not well-designed error if the input was not an iterable. #446
  • Fixed image normalization crashing when an input ndarray of multiple images was changed during augmentation to a list of multiple images with different shapes and the original input ndarray represented a single image or a collection of 2D (H,W) images. This problem affected augment(), augment_batch() and augment_batches().
  • Fixed a typo in an image normalization error message. #451
  • Fixed a problem in WithChannels that could lead random sampling in child augmenters being unaligned between images and corresponding non-image data. #451
  • Added aliases to imgaug.random.RNG for some outdated numpy random number sampling methods that existed in numpy.random.RandomState but not in numpy's new RNG system (1.17+). These old methods are not used in imgaug, but some custom augmenters and Lambda calls may require them when interacting with the provided random_state instances. #486
  • Fixed Affine producing unaligned augmentations between images and segmentation maps or heatmaps when using translate_px and the segmentation map or heatmap had a different height/width than corresponding image. #489
  • Fixed a crash in SnowflakesLayer that could occur when using values close to 1.0 for flake_size. #471
  • Fixed MultiplyHueAndSaturation crashing if the RNG provided via random_state was not None or imgaug.random.RNG. #493
  • Fixed CloudLayer.draw_on_image() producing tuples instead of arrays as output for float input images. #540
  • Fixed Affine parameter translate_px behaving like translate_percent if a continuous stochastic parameter was provided. Analogously translate_percent would behave like translate_px if a discrete stochastic parameter was provided. #508
  • Fixed code hanging indefinitely when using multicore augmentation on NixOS. #414 #510
  • Fixed a deprecation warning and potential crash in python 3.8 related to the use of collections instead of collections.abc. #527
  • Fixed deprecated scipy.fromfunction() being called. #529
  • Fixed imgaug.random.normalize_generator() crashing in numpy 1.18. The function relied on numpy.random.bit_generator.BitGenerator, which was moved in numpy 1.18 to numpy.random.BitGenerator without a deprecation period for the old name. #534
  • Fixed an issue that could lead to endlessly hanging programs on some OS when using multicore augmentation (e.g. via pool) and augmenters using OpenCV. #535
  • Fixed imgaug.random.seed() not seeding the global RNG in-place in numpy 1.17+. The (unfixed) function instead created a new global RNG with the given seed. This set the seed of augmenters created after the seed() call, but not of augmenters created before the seed() call as they would continue to use the old global RNG. #557
  • Fixed cval in ElasticTransformation resulting in new pixels in RGB images being filled with (cval, 0, 0) instead of (cval, cval, cval). #561 #562
  • Fixed some augmenters in module weather not transferring seed values or random states that were provided upon creation to child augmenters. #568
  • Fixed an inaccuracy in PerspectiveTransform that could lead to slightly misaligned transformations between images and coordinate-based augmentables (e.g. bounding boxes). The problem was more significant the smaller the images and larger the scale values were. It was also worsened by using fit_output. #585
  • Fixed KeepSizeByResize potentially crashing if a single numpy array was provided as the input for an iterable of images (as opposed to a list of numpy arrays). #590