Image classification dataset containing 1360 images consisting of 17 species of flowers (trimaps/bounding box annotations only available for 16):

• Source (README)

• Original images (60MB), original trimaps (60MB)

• Image classification dataset (35MB)

• Object detection dataset (VOC XML format) (61MB - not all images have annotations)

• Object detection dataset (ADAMS format) (35MB)

Conversion from original data:

• Decompress images and trimaps archives

• Create directory voc

• Copy all images from jpg directory into voc

• Run Python script (requirements.txt) from the directory with jpg and voc directories to generate the bounding boxes in VOC XML format

• The voc directory contains the object detection dataset

• Run: wai-annotations convert from-voc-od -i "voc/*.xml" od-to-ic to-subdir-ic -o subdir (requires the wai.annotations library; tested with 0.7.5)

• The subdir directory contains the image classification dataset

License

Unclear.