Low rank sparse coefficient based nuchal translucency image de-noising

Abstract

The process of eliminating distortion or noise from an image is known as image de-noising. Random noise is introduced to ultrasonic imaging, resulting in reduced contrast in the images. For Nuchal translucency (NT) detection, image de-noising is a crucial stage. Although deep-learning methods have been extensively studied for this problem and have shown compelling results, most networks may result in disappearing or inflating gradients and need more memory and time to attain a spectacular performance. To achieve better overall framework optimization, Novel Methodology of anisotropic filtering followed by a compressed sensing based on Low Rank Sparse Coefficient(LRSC) for ultrasound image de-noising is proposed to achieve better overall framework optimization, this article proposes anisotropic filtering followed by a compressed sensing based on Low Rank Sparse Coefficient(LRSC) for ultrasound image de-noising. This hybrid technique is quite effective at reducing noise while yet retaining fine image details. Real-time hospital images are utilized to assess the efficiency of the proposed model, taking into account clinical accessibility and imaging features. The peak signal-to-noise ratio (PSNR), mean square error (MSE), and structural similarity index (SSIM) were used to evaluate the proposed method’s performance. Average SSIM, PSNR, MSE values are 0.98, 42.28 and 49 for HCNN[12], GAN[16] and Proposed method respectively. Proposed method have average MSE of 2.6, HCNN have 14 and GAN have 371.