IMAGE COMPRESSION OF RADIOGRAPH USING NEURAL NETWORK AND WAVELET

Bandwidth conservation is an important issue in case of multimedia communication. Uncompressed multimedia (graphical, audio and video) data requires considerable storage capacity and transmission bandwidth. Despite rapid progress in mass-storage density, processor speeds and digital communication system performance, it demands for data storage capacity and data-transmission bandwidth continuously to outstrip the capabilities of available technologies. So to solve this problem an efficient multimedia communication scheme is proposedwhich is based on Wavelet. Image compression is the technique of reducing the size of image file without degrading the quality of the image. There are many techniques available in the lossy image compression in which Wavelet transform based image compression is the best technique. Various types of Wavelets are used for image compression. This paper shows Better image compression by using different wavelet with the help of Neural network. The paper defines the progress made towards calculating different parameter for Wavelet and after that determines the wavelet which gives minimum value of mean square error and maximum value of peak signal to noise ratio. By this best compression Wavelet is obtained. For Analysis considered MSE value should be a minimum and peak signal to noise ratio value should be a maximum. By implementing neural network, the optimum image compression system use a supervised neural network based on the back propagation learning algorithm, due to its implementation, simplicity and the availability of sufficient target database for training the supervised learner is obtained. The paper present the idea of image compression based on hierarchical back propagation neural network and results are analyzed. The further analysis is conducted in the network model and tested training algorithm. Finally image compression and image reconstruction are accomplished respectively, a minimum accuracy of 89% was considered as accepted. The neural network yielded 98.65% correct recognition rate ofoptimum compression ratios, This concludes that a high compression ratio is achieved with Bi-orthogonal Wavelet functions. Theresults are obtained with a Bi-orthogonal 6.8 Reconstruction Wavelet function and proved the best. Then Neural Network is implemented to prove the best result and hence achieved. Experimental results suggest that the proposed system can be efficiently used to compress while maintaining high image compression.