image-dataset-converter 0.0.4

Python3 library for converting between various image annotation dataset formats.

The image-dataset-converter library allows the conversion between various dataset formats for image annotation datasets. Filters can be supplied as well, e.g., for cleaning up the data.

Dataset formats:

• image classification: ADAMS (r/w), sub-dir (r/w)

• object detection: ADAMS (r/w), COCO (r/w), OPEX (r/w), ROI (r/w), VOC (r/w), YOLO (r/w)

• image segmentation: blue-channel (r/w), grayscale (r/w), indexed PNG (r/w), layer segments (r/w)

Changelog

0.0.4 (2024-07-16)

• limiting numpy to <2.0.0 due to problems with imgaug library

0.0.3 (2024-07-02)

• switched to the fast-opex library

• helper method from_indexedpng was using incorrect label index (off by 1)

• Data.save_image method now ensures that source/target files exist before calling os.path.samefile

• requiring seppl>=0.2.6 now

• readers now support default globs, allowing the user to just specify directories as input (and the default glob gets appended)

• the to-yolo-od writer now has an option for predefined labels (for enforcing label order)

• the to-yolo-od writer now stores the labels/labels_cvs files in the respective output folders rather than using an absolute file name

• the bluechannel/grayscale/indexed-png image segmentation readers/writers can use a value other than 0 now for the background

• split filter has been renamed to split-records

0.0.2 (2024-06-13)

• added generic plugins that take user Python functions: from-pyfunc, pyfunc-filter, to-pyfunc

• added idc-exec tool that uses generator to produce variable/value pairs that are used to expand the provided pipeline template which then gets executed

• added polygon-simplifier filter for reducing number of points in polygons

• moved several geometry/image related functions from imgaug library into core library to avoid duplication

• added python-image-complete as dependency

• the ImageData class now uses the python-image-complete library to determine the file format rather than loading the image into memory in order to determine that

• the convert-image-format filter now correctly creates a new container with the converted image data

• the to-coco-od writer only allows sorting of categories when using predefined categories now

• the from-opex-od reader now handles absent meta-data correctly

• added the AnnotationsOnlyWriter mixin for writers that can skip the base image and just output the annotations

0.0.1 (2024-05-06)

• initial release