From my earlier work on vision processing for people with low or no sight.
- Learning scene representations for human-assistive displays — self-attention networks for prosthetic vision pipelines
- Finding the vanishing point in images — Canny, Hough, and a bit of geometry