Automatic brain MR image denoising based on texture feature-based artificial neural networks

Abstract

Abstract. Noise is one of the main sources of quality deterioration not only for visual inspection but also in computerized processing in brain magnetic resonance (MR) image analysis such as tissue classification, segmentation and registration. Accordingly, noise removal in brain MR images is important for a wide variety of subsequent processing applications. However, most existing denoising algorithms require laborious tuning of parameters that are often sensitive to specific image features and textures. Automation of these parameters through artificial intelligence techniques will be highly beneficial. In the present study, an artificial neural network associated with image texture feature analysis is proposed to establish a predictable parameter model and automate the denoising procedure. In the proposed approach, a total of 83 image attributes were extracted based on four categories: 1) Basic image statistics. 2) Gray-level co-occurrence matrix (GLCM). 3) Gray-level run-length matrix (GLRLM) and 4) Tamura texture features. To obtain the ranking of discrimination in these texture features, a paired-samples t-test was applied to each individual image feature computed in every image. Subsequently, the sequential forward selection (SFS) method was used to select the best texture features according to the ranking of discrimination. The selected optimal features were further incorporated into a back propagation neural network to establish a predictable parameter model. A wide variety of MR images with various scenarios were adopted to evaluate the performance of the proposed framework. Experimental results indicated that this new automation system accurately predicted the bilateral filtering parameters and effectively removed the noise in a number of MR images. Comparing to the manually tuned filtering process, our approach not only produced better denoised results but also saved significant processing time.

CUDA-based acceleration and BPN-assisted automation of bilateral filtering for brain MR image restoration

“…Among 26 candidate image texture features studied in our earlier work, this article exhaustively investigated the optimal texture features based on the SFFS algorithm associated with the BPN scheme. In contrast to the nine texture features obtained in our preceding study, the outcome was six optimal and independent texture features. They consisted of Mean belonging to the fundamental statistics, CON being the Tamura texture feature, and four (RLN(135°), RP(135°), LGRE(45°), LGRE(90°)) in the GLRLM category.…”

“…In our earlier study, we have investigated the correlations between the bilateral filter parameters and the image texture features, which included the fundamental statistics, gray‐level co‐occurrence matrix (GLCM), gray‐level run‐length matrix (GLRLM), and Tamura texture features. By applying the paired‐samples t‐test, 26 texture features were selected as the candidate set according to the average p ‐value as summarized in Table .…”

“…Essentially, the sequential forward floating selection (SFFS) is an adequate algorithm to simultaneously address these problems. The SFFS is a bottom up optimal search procedure that overcomes the “nesting effect” problem of the sequential forward selection (SFS) method, which was utilized in our previous investigation . The nesting effect means that the features once selected cannot be discarded, leading to suboptimal results in the SFS.…”

“…In our previous study, we demonstrated that some image texture features revealed typical regularities of biological structures in brain MR images and were advantageous to the automatic parameter decision procedure of the bilateral filter. However, in that study, the optimal combination of texture features, the filter parameter estimation accuracy, and the corresponding denoising performance were not thoroughly explored.…”

“…This article is in an attempt to intensively investigate the mechanism for achieving fully automatic brain MR image restoration with an accelerated and parameter‐free bilateral filter. Comparing to our earlier investigation, we additionally and uniquely exploit the GPU‐based acceleration of the traditional bilateral filter for efficient computation. Optimization techniques to expedite the computation in either thread usages or memory resources are addressed.…”