Adaptive Pixel-Based Data Fusion for Boundary Detection

It is of practical importance to fuse data obtained by multiple sensors for improving the performance of computer vision systems. This paper introduces an algorithm for pixel-based data fusion on the variational framework. An adaptive system fuses data eeectively using a variational technique. Previously, we have introduced a technique to fuse gray-scale image and texture extracting features for segmenting an image with both textured and non-textured surfaces. This paper extends the study for more general multi-valued data and improve the previous algorithm in terms of the performance and speed.