Why this page exists. Every claim the site makes about the corpus is a claim about individual photographs, so every photograph should be linkable on its own. This is the permanent record for one specimen: the image, the six feature scores against the written rubric, the three hypothesis readings, and its identification status, stated plainly even when the honest answer is unknown.
Scores were assigned image by image against the written rubric; the full matrix and methodology are documented in the project knowledgebase, and the complete dataset is a CSV download. To test the scoring yourself without seeing the answers first, use the blind scoring page.