Selective Tuning Visual Attention Model
J. Amudha · 2009
In this paper, a biologically motivated attention system based approach is introduced which detects regions of interest in images based on concepts of the Human visual system. The pre-attentive stage of constructing the feature maps uses low-level visual features such as color and intensity that are computed as individual channels by incorporating task parallelism. The selective tuning model selects the most salient region. The system copes with the requirements of human attention model and designed for real time application. Visual attention is a mechanism in human perception, which selects salient regions from a scene and provides these regions for higher-level processing as object recognition. Computational vision systems (8),(7) face the same problem as humans due to the large amount of available information that has to be processed. For an efficient processing the order in which a scene is investigated has to be determined in an intelligent way. A promising approach to achieve this is the use of computational attention systems that simulate human visual attention. Attention is the process of selecting and gating visual information based on saliency in the image itself (bottom-up), and on prior knowledge about scenes, objects and their interrelations (top-down). Visual attention addresses both problems by selectively enhancing perception at the attended location, and by successively shifting the focus of attention to multiple locations. Proposed visual attention system is based on the Itti et al. (6) implementation of the saliency-based model of bottom-up attention. Section II discusses the related visual attention models and section III explains the extensions in the same formal framework. In section IV the analysis of the performance of the model has been tested with various cases and compared with Itti's saliency tool box. In section V conclusions and further enhancements are discussed.