Although many lung disease diagnostic procedures can benefit from computer-aided detection (CAD), current CAD\r\nsystems are mainly designed for lung nodule detection. In this article, we focus on tuberculosis (TB) cavity\r\ndetection because of its highly infectious nature. Infectious TB, such as adult-type pulmonary TB (APTB) and\r\nHIV-related TB, continues to be a public health problem of global proportion, especially in the developing countries.\r\nCavities in the upper lung zone provide a useful cue to radiologists for potential infectious TB. However, the\r\nsuperimposed anatomical structures in the lung field hinder effective identification of these cavities. In order to\r\naddress the deficiency of existing computer-aided TB cavity detection methods, we propose an efficient\r\ncoarse-to-fine dual scale technique for cavity detection in chest radiographs. Gaussian-based matching, local binary\r\npattern, and gradient orientation features are applied at the coarse scale, while circularity, gradient inverse\r\ncoefficient of variation and Kullbackââ?¬â??Leibler divergence measures are applied at the fine scale. Experimental results\r\ndemonstrate that the proposed technique outperforms other existing techniques with respect to true cavity\r\ndetection rate and segmentation accuracy.
Loading....