Small objects detection in image data based on probabilistic visual learning

In this paper, we present an efficient appearance-based method for the detection of small objects in images. In the detection scheme, the probabilistic visual learning (PVL) technique is used for modeling the appearance of small objects and constructing a saliency measure function. Based on this function and the feature vector extracted at each pixel position, a small object saliency map is formed by lexicographically scanning the input image. We treat such saliency map as a spatially filtered result of input image. Compared to several filter-based detection methods, experiments show that the proposed algorithm outperforms these methods.