Automatic Target Recognition from IR Images using Bottom-Up Selective Attention

摘要

An automatic target recognition (ATR) system is implemented using visual selective attention. Typical ATR system has 3 stages, which are target detection, clutter rejection, and target recognition. IR tank images are used for testing performance of the ATR system. In target detection stage, "Saliency map" for bottom-up selective attention, and multilayer perceptron (MLP) are used to extract features at local area. And we combine these two maps with another MLP. A clutter rejection stage and a target recognition stage are implemented using MLP. The performance of each stage and also overall performance is compared for three target detection methods. When two maps are combined using MLP, not only target detection rate but also overall performance is enhanced. Proposed method shows high recognition rate and low false alarm.