Tuberculosis malady recognition in chest radiographs via artificial neural networks

摘要: This proposed method will give a suggestion for Tuberculosis (TB) diagnosing using Artificial Neural Networks (ANN). Since diagnostic imaging techniques such as x-rays (Radiographs), Magnetic Resonance Imaging (MRI), Computed Tomography (CT) are available, X-ray is widely preferred edging the image of TB affected area in chest region. due to its fastness and easy also inexpensive nature. The couldn't show small cancer blood clot chest, it works fine categorize TB. But accuracy disease diagnostics depends on ability practitioner owing lack automatic approach. Hence this paper focuses present an automated approach recognize conventional posteroanterior Radiographs. includes 3 stages (i) segmentation lung region from radiographs. (ii) Extract set features (iii) classification section identify presence absence Initially Graph-cut used algorithm extract Region Interest (ROI) input then be done by Network (ANN) combination with Genetic Algorithm (GA). ANN utilizes Levenberg-Marquardt (LMA) train neuron.