AccessLens Datasets
AccessDB/AccessReal (COCO annotation format)
- AccessDB contains 2,388 indoor scene images from ADE20K dataset, re-annotated with inaccessibility classes.
- AccessReal contains 42 high-resolution real-world indoor images for evaluation of the detector trained with AccessDB.
- labels.txt presents 21 inaccessibility classes and 1 'unidentifiable' class (marked as 'unknown' in the dataset)
3D Assistive Augmentation Dictionary
- 3A2 dictionary contains 280 3D-printed augmentations for 52 everyday objects (e.g., handle, door, knob, book, nail clipper, knife, hair dryer, microwave, stove, table, etc.) with potential inaccessibility context, fully annotated with AccessMeta categories (actuation, constraint, indication).