Abstract

-Region of interest (ROI) extraction is an important step in deriving visual features for an audio-visual speech recognition system. Colour based segmentation offers the potential of computationally inexpensive algorithms for ROI selection. This paper presents a comparative study of two colour based techniques, one using hue and accumulated difference, the other chrominance. Results are presented for the CUAVE database. The two methods achieved 69% and 72% correct ROI extraction. The experiment prompted investigation of a new method using a chrominance based accumulated difference image. The new method achieved 79% correct ROI identification. The overall results suggest that a dual approach using chrominance to locate the mouth region and only employing an accumulated difference image when significant motion is not present would offer good robustness with lower computational cost.

Mean squared error (MSE) and peak signal-to-noiseratio (PSNR) are the most common methods for measuring the quality of compressed images, despite the fact that their inadequacies have long been recognized. Quality for compressed still images is occasionally evaluated using human observers who provide subjective ratings of the images. Both SNR and subjective quality judgments, howevel; may be inappropriate for evaluating progressive compression methods which are to be used for fast browsing applications. In this paper; we present a novel experimental and statistical framework for comparing progressive coders. The comparisons use response time studies in which human observers view a series ofprogressive transmissions, and respond to questions about the images as they become recognizable. We describe the framework and apply it to the comparison of several well known progressive algorithms. ~~ ~ 'This work was supported by the National Science Foundation under Grants MIP-9617366 and MIP-9624729 (CAREER), and by the Center for Wireless Communications at UCSD.