Computer Aided Classification of X-ray Images with MLP Based BPNN Method with Extracted Texture Features
With advancement of new technology, computerized classification of X-ray images plays an important role in giving faster results and very accurate results. Classification of X-rays is an initial step for fracture detection of x-ray images. In many rural areas where there are no proper hospitals and medical facilities and that too, there may not find sufficient orthopeadicians for immediate treatment for fractures. As per the reports of WHO accidents are the 2nd largest cause for deaths which occurs majorly with fractures in the bones. In many government hospitals, sufficient orthopedics is not available. In order to overcome this a computer based detection plays a major rolein which classification becomes the first step.Proposed system involves totally 5 steps. denoising of images using M3 filter,Image enhancement using adaptive graph,extraction of texture features and last categorization of images using neural network.Here for classification,the images of leg,foot,chest,spine,neck,hand are taken.RBFNN(Radial Basis Function Neural Network) and MLP(multi layer perception) based neural network and SVM(Support Vector Machine) are used for classifying the X-ray images.Results shows that MLP based neural network gives more accuracy of 90% when compared to RBFNN and SVM.Hence we consider here MLP based NN is used as efficient tool for classification.
Keywords: Radial basis function neural network(RBFNN), Texture feature extraction, Multi layer based backpropagation neural network(MLP-BPNN), Support Vector Machine (SVM).