Data Descriptor

An automated metric or feedback mechanism that characterizes the quality or properties of a dataset, particularly training images used in machine learning. In accessibility research, data descriptors provide non-visual feedback to blind users about the quality of photos they have taken to train object recognizers — for example, indicating whether an object is cropped out of frame, too small, blurred, occluded by a hand, or lacking variation across the training set. Data descriptors bridge the gap between visual verification (which sighted users rely on) and accessible alternatives that allow blind users to independently assess and improve their training data.