A novel multistage system for the detection and removal of pectoral muscles in mammograms

Abstract

Abstract:In this paper, a novel multistage scheme for pectoral muscle removal from mammography images is proposed, and the performance of this system is verified using the publicly available Mammographic Image Analysis Society digital mammogram database. This database is composed of mediolateral oblique mammography images including three different tissue types (fatty, fatty-glandular, and dense-glandular) with three health status types (normal, benign cancer, and malignant cancer). In the implementation of the proposed system, a mammography image is first preprocessed by performing noise reduction background removal followed by artifact suppression processes. Then a presegmentation procedure is applied using region growing and line fitting is executed. Finally, pixels including pectoral muscle regions are removed from mammography images with an accuracy of 94.40%, sensitivity of 89.62%, and specificity of 99.99% after some postprocessing operations. Although the mean false positive rate obtained by the proposed approach is higher than that of other studies in the literature, not only the lower mean false negative rate but also the enhancement in the quality of pectoral muscle segmentation for all 322 images in the database evidently show the success of the proposed approach.

A Hybrid Textural and Geometrical Feature Extraction to Reveal Hidden Information From Suspicious Regions on Mammograms

“…In this paper, textural and geometrical features are utilized on a pre-designed CAD system [34,35] in the light of a radiologist-eye, for breast cancer diagnosis. The experiments are performed on a computer with I5− 7200U at 2.5 GHz and 8-Gb memory and have been carried out using MATLAB R2021a.…”

“…Initially, all mammography images are resized to a size of 256×256 using bi-cubic interpolation, for ease of operation, and adaptive median filter is applied on mammography images for digitization noise reduction while preserving gross details such that pectoral muscle and ROI edges [34]. Then, labelling artifacts are suppressed and the background is removed by defining the breast parenchyma as the largest area of the binarized noise-reduced mammograms [34]. As the last step of the pre-processing stage, pectoral muscle removal is realized using a region growing algorithm [34].…”

“…Then, labelling artifacts are suppressed and the background is removed by defining the breast parenchyma as the largest area of the binarized noise-reduced mammograms [34]. As the last step of the pre-processing stage, pectoral muscle removal is realized using a region growing algorithm [34]. Figure 3 demonstrates the pre-processing stage on a sample image [35].…”

“…Under these circumstances, textural, and geometrical feature extraction techniques such as Centroid Distance (CD) approach [29] and Edge Step (ES) approach [30] are thought to be descriptive for health status. Accordingly, feature vectors defining ROIs' textural and geometrical structures are examined via a pre-designed CAD system [34,35] for breast cancer diagnosis, in this paper. Noise reduction, background removal, artifact suppression, and pectoral muscle removal are carried out [34] and the ROIs are identified [22] in the pre-processing and detection stages of this CAD system, respectively.…”

“…Accordingly, feature vectors defining ROIs' textural and geometrical structures are examined via a pre-designed CAD system [34,35] for breast cancer diagnosis, in this paper. Noise reduction, background removal, artifact suppression, and pectoral muscle removal are carried out [34] and the ROIs are identified [22] in the pre-processing and detection stages of this CAD system, respectively. In the feature extraction stage, an Adaptive Convex Hull (ACH) approach is examined for geometrical feature extraction while textural formation of ROIs is formalized by Haralick features [36][37][38] computed from Gray-Level Co-Occurrence Matrix (GLCM) of mammogram ROIs.…”