Information Fusion for Image Analysis: Neural Methods and Technology Development

摘要

Research at Boston University has produced computational models of biological vision and learning that embody a growing corpus of scientific data and predictions. Vision models perform long-range grouping and figure/ground segmentation, and memory models create attentionally controlled recognition codes that intrinsically combine bottom-up activation and top-down learned expectations. These two streams of research form the foundation of completed projects that define novel dynamically integrated systems for image understanding. Simulations using multi-spectral images illustrate road completion across occlusions in a cluttered scene, information fusion from input labels that are simultaneously inconsistent and correct, and applications of models of color vision. The CNS Technology Lab has further integrated science and technology through analysis, testing, and development of cognitive and neural models for large-scale applications, complemented by software specification and code distribution.