Project Images
Learned basis functions for natural scenes. The ICA transformation learns a set of basis functions where each basis function is localized in space and oriented similar to receptive fields in the visual cortex V1.

Project Description
The efficient encoding of visual sensory information is an important task for image processing systems and its study may provide insights into coding principles of biological visual systems. We are interested in finding efficient codes for image processing problems such as image classification, segmentation and image compression.

Project Goals
- Image feature extraction by unsupervised learning
- Image classification and segmentation
- Finding efficient codes for color

Schedule
.

Phase Name End Date Phase 1 July 15, 2001

Staff
Suman Mitra
Eizaburo Doi
.