Local Spatio-Temporal Analysis in Vision Systems:

摘要

This is the final progress report of teh vision group at the University of Texas under support of AFOSR URI grant F4962O-93-1-O3O7. In this report we will attempt to summarize the major accomplishments over the previous 5 years. Aim 1: to develop a mathematical model of the initial stages of visual processing (the front end mechanisms), based upon a wide range of physiological and psychophysical data. Aim 2: To develop new methods and models of local frequency coding. Aim 3: To develop new mathematical models and computer-vision algorithms for performing complex visual tasks that are based upon local frequency coding representations. Aim 4: To develop models for human performance in complex visual tasks that build upon current understanding of the front-end mechanisms. Aim 5: To develop a computational testbed for implementing, comparing, integrating and visualizing the different models and modules developed during the project, using a massively parallel machine and graphics workstation front-end.