Further Development of a Computational Vision Model for the Generation and Analysis of Photo-Realistic Scenes for Military and Civilian Virtual Reality Applications

摘要

This document and the accompanying software constitute CDRL A002, Final Report, CDRL A003, Computer Software, and CDRL A004, Software User's Manual, for Small Business Innovative Research Phase 1 contract number DAAE07- 95-C-R071, entitled 'Further Development of a Computational Vision Model for the Generation and Analysis of Photo-Realistic Scenes for Military and Civilian Applications.' The objective of the Phase 1 effort was to investigate the application of computational vision models (CVM) to simulate human performance in image analysis tasks, military target acquisition, automotive collision avoidance, and visual virtual reality for perception testing. The specific objectives are listed below: (1) To develop, implement and assess algorithms and software to generate synthetic imagery with photo-realistic appearance of the natural surrounding image, for application to create low signature stimuli for visual virtual reality for perception testing; (2) To define functional requirements to enhance and refine the computational vision model of the 'early' stages of visual processing (from entrance at the pupil through receptive field linear spatial filtering) implemented in the TARDEC National Automotive Center Visual Performance Model (VPM); and (3) To develop and implement a model of alternative hypotheses regarding 'late' vision processing stages (i.e., alter the 'early' stages of visual processing, up to cognitive decision and response) for the purpose of formulating alternative signature metrics to predict decision and response behavior.