De-Speckling Breast Cancer Ultrasound Images Using a Rotationally Invariant Block Matching Based Non-Local Means (RIBM-NLM) Method

Abstract

The ultrasonic technique is an indispensable imaging modality for the diagnosis of breast cancer in young women due to its ability to efficiently capture the tissue properties and decrease the negative recognition rate, thereby avoiding non-essential biopsies. Despite the advantages, ultrasound images are affected by speckle noise, generating fine-false structures that decrease the contrast of the images and diminish the actual boundaries of tissues in the ultrasound image. Moreover, speckle noise negatively impacts the subsequent stages of the image processing pipeline, such as edge detection, segmentation, feature extraction, and classification. Previous studies have formulated various speckle reduction methods for ultrasound images; however, these methods suffer from being unable to retain finer edge details; they also require more processing time. In this study, we propose a breast ultrasound de-speckling method based on rotationally invariant block matching non-local means (RIBM-NLM) filtering. The effectiveness of our method is demonstrated by comparing our results with those for three established de-speckling techniques: the switching bilateral filter (SBF), the non-local means filter (NLMF), and the optimized non-local means filter (ONLMF). We analyzed 250 images from a public dataset and six images from a private dataset. Evaluation metrics, including the Self-Similarity Index Measure (SSIM), the Peak Signal-to-Noise Ratio (PSNR), and the Mean Square Error (MSE), were utilized to measure performance. With the proposed method, we were able to record an average SSIM of 0.8915, a PSNR of 65.97, an MSE of 0.014, an RMSE of 0.119, and a computational speed of 82 s at a noise variance of 20 dB using the public dataset, all with a p-value of less than 0.001 compared against NLMF, ONLMF, and SBF. Similarly, the proposed method achieved an average SSIM of 0.83, a PSNR of 66.26, an MSE of 0.015, an RMSE of 0.124, and a computational speed of 83 s at a noise variance of 20 dB using the private dataset, all with a p-value of less than 0.001 compared against NLMF, ONLMF, and SBF.

Graph neural network-based breast cancer diagnosis using ultrasound images with optimized graph construction integrating the medically significant features

“…The noteworthy performance difference is occurred for several reasons. Firstly, the raw breast ultrasound images contain high amount of speckle noise which results poor visualization and tumor tissue boundary minimization (Ayana et al 2022 ). Therefore, the raw image may contain redundant features which can mislead the classification model extensively.…”

Improving the Automated Diagnosis of Breast Cancer with Mesh Reconstruction of Ultrasound Images Incorporating 3D Mesh Features and a Graph Attention Network

“…Speckle noise in ultrasonic images reduces tissue boundaries, produces deceptive structures, and makes it difficult to segment and classify them. For which the tumor's depth, edges, and pixel distributions cannot be identified in some cases [33]. Constructing a 3D view from an image that lacks comprehensive shape and edge information is a complex task as it needs to predict the incomplete parts effectively [34][35][36].…”

Pathological Insights: Enhanced Vision Transformers for the Early Detection of Colorectal Cancer

“…Moreover, most state-of-the-art deep learning studies for early detection of CRC are solely based on convolutional neural networks (CNNs) [28][29][30]. CNNs can learn visual representations for easy transfer and strong performance, owing to the strong inductive bias of spatial equivariance and translational invariance provided by their convolutional layers [31][32][33][34][35]. However, vision transformers (ViTs) exhibit superior performance over CNNs for natural image classification and segmentation [36][37][38][39].…”