Stereo Processing in the Human Visual System
These two videos describe some key aspects of human stereo vision and introduce a computational model that tries to incorporate these observations. The first video addresses the features used for stereo matching and the important role that multiple scales of processing and vergence eye movements play in human stereo vision. It also briefly touches on the neural mechanisms underlying stereo vision in monkeys. The second video describes a stereo correspondence algorithm proposed by Marr, Poggio, and Grimson that is based on matching zero-crossings of images processed with Laplacian-of-Gaussian operators of different size.
Videos
The following links take you to the videos on YouTube:
- Human Stereo Vision (24:36)
- Marr-Poggio-Grimson Stereo Algorithm (33:04)
The video files are also stored in our shared folder on Google Drive, named
human_stereo_vision.mp4 (537 MB) and MPG_stereo_algorithm.mp4 (850 MB).
Video Transcripts
The following pdf files provide transcripts of the videos:
The above pdf files are also stored in the shared folder on Google Drive.
Slides
The following pdf files provide the slides presented in the videos:
The above pdf files are also stored in the shared folder on Google Drive.