Removing the Blurring from X-Ray Image Using BM3D Technique

The x-ray image quality of normal patient is needed to enhance to diagnose accurately. For this reason, block-matching 3D (BM3D) technique is chosen for denoising the x-ray images. The currently the best BM3D denoising system utilizes a white Gaussian noise (WGN) design. The similar 2D x-ray image is converted to 3D data arrays by grouping to improve the sparsity and it is called grouping. Collaborative filtering is a unique method for dealing with these three-dimensional groups. The collaborative filtering reduces noise, demonstrating even the details of image shared by grouped blocks while preserving the crucial unique characteristics from every individual block. After that, the shifted blocks are replaced with new positions. As these blocks coincide, we get a variety of special predictions with each pixel, which we have to combine. The Wiener filtering process is implemented in the transform coefficients to a post-thresholding signal in the present BM3D algorithm for improved noise removal. Wiener filtering of transform domain co-efficient is used based on the properties of x-ray images in terms of PSNR and SNR value. The hard thresholding system is used in previous step to denoise the x-ray image in the utter lack of a ground-truth signal. The performance of BM3D technique is compared with wavelet transform to evaluate image quality. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.