An Efficient Combined Approach For Denoising Fibrous Dysplasia Images
A. Saranya, K. Kottilingam · 2021 International Conference on System, Computation, Automation and Networking (ICSCAN) · 2021
Image denoising and Reconstruction are necessary for medical image processing. It helps to crate a lot of research scope in the medical image analytics. The proposed model of this paper is a combined approach for reducing the noisy details from the bone images. It also predicts the Peak Signal to Noise Ratio (PSNR) ratio between the quality and noise. This approach is greater idea to track the fibrous tissue growth in the bones. This framework is the combination of Auto-Encoder (AE) and Convolutional Neural Network (CNN). AutoEncoder is used for denoising the image with higher quality. Enhanced Convolutional Neural Network predicts the minor noisy details in the image with deep hidden layers. The convolution algorithm creates the major optimism for reconstruction the highly corrupted images. Similarly this framework helps to process denoising and extracting the important features from the image for grouping the regions. Our proposed model achieves the PSNR range for denoised image is 65.369 db.