Image-object-slicer

New
- Add support to the YOLO annotation format by adding `YOLOParser`.

Improved
- Simplify `split_file` in all `SingleFileAnnotationParser` parsers to generate items more efficiently and in sequence;
- Remove `chunksize` argument from all calls to `imap_unordered`;
- Replace list comprehensions with generator comprehensions as iterators for `imap_unordered`;
- Turn all `split_file` methods into generators.

Fixed
- Remove support to the MOT annotation format by removing `MOTParser`;
- Fix image names and globs when files are in a subfolder;
- Default label to `face` when not available in `parse_item` in `WIDERFaceParser`;
- Only find labels file if any other annotation file was found;
- Raise exception if no labels file was found.

New
- Add support to the MOT annotation format by adding `MOTParser`;
- Add `labels` and `parse_labels` to `SingleFileAnnotationParser` and `MultipleFileAnnotationParser`.

Improved
- Optimize `split_file` in `COCOParser`;
- Use `csv` in `parse_file` in `KITTIParser`.

New
- Add support to the Open Images annotation format by adding `OpenImagesParser`.

Improved
- Use `int` for absolute coordinates (in pixels) and `float` for coordinates relative to size of image in item dicts.

New
- Add support to the WIDER Face annotation format by adding `WIDERFaceParser`.

Improved
- Remove empty items in `split_file`.

New
- Add support to the KITTI annotation format by adding `KITTIParser`.

Improved
- Remove empty items in `split_file` of `COCOParser`;
- Add instructions to install from PyPI.

Fixed
- Fix glob of `COCOParser`;
- Fix issue with PyPI.

New
- Add support to the MS COCO Object Detection annotation format by adding `COCOParser`.