Hybrid Denoising Development to Improve the Quality of Image Segmentation with Noise

Image noise can reduce image quality and loss of some detailed image information. Therefore, a noise reduction process called image denoising is needed so that the image quality becomes better. However, the main problem with denoising is how to eliminate noise while preserving edge. The denoising process also aims to improve the quality of segmentation because the image to be processed has better quality. This study aims to build an effective method so that the denoising process maintains detail and edges to eliminate random-valued noise (gaussian, and speckle noise) and fixed-valued noise (salt & pepper noise) by performing Bilateral Filtering and Non-Local Means Filtering on the approximation coefficient of the Discrete Wavelet Transform (DWT). Whereas for the detail coefficient, Soft-Thresholding will be carried out and followed by the Anisotropic Diffusion Filtering process. Trials are carried out with grayscale images added with noise. The results show that the proposed method produces a higher average PSNR than the other methods, which is equal to 29,26 db. Next, the segmentation applying Fuzzy C-Means with spatial correlation is carried out. This method utilizes spatial information based on the correlation values between pixels to modify the weights for computing the pixel memberships and the cluster centroids of the basic FCM. The modified FCM is initialized using the result of the basic FCM. The segmentation performance is represented by the sensitivity, specificity, and accuracy between the segmented image and ground truth.