Transcript and context
A typical demo: a well-known artwork, photographed. The user takes a picture of the painting and asks Lumyeye to describe it. The AI replies: "You're looking at Van Gogh's The Starry Night. A swirling blue and yellow sky over a sleeping village, with a dark cypress in the foreground." A painterly, sensory description that captures the intent of the work, not a flat inventory of elements.
For museum visits this is especially valuable. Lumyeye recognizes a large catalog of referenced works (painting, sculpture, art photography) and delivers both a visual description and cultural context (artist, date, movement). Captions next to the artwork are also OCR'd if the user points at them.
The same usage extends to personal photos. A family album photographed page by page, a souvenir picture, a postcard received in the mail: Lumyeye describes the scene, identifies people (without naming them, to respect privacy), describes facial expressions, clothing, surroundings. The user regains access to their visual memory.
The description is VoiceOver-friendly. Users who already have their iOS accessibility setup in place can adopt Lumyeye without disrupting it. Lumyeye's voice layers cleanly on top of VoiceOver, and the two switch between each other without conflict.