Open AccessEffective Strain Image Sequence Selection by Using Semi-Automated Image Processing Technique
Author(s) -
Gulam Mahfuz Chowdhury,
Md. Mahedi Hasan,
Asif Ahmed,
Md. Wahid Tousif Rahman,
Md. Taslim Reza
Publication year - 2021
Publication title -
gub journal of science and engineering
Language(s) - English
Resource type - Journals
ISSN - 2409-0476
DOI - 10.3329/gubjse.v7i0.54022
Subject(s) - pixel , visibility , artificial intelligence , gray level , computer science , computer vision , sequence (biology) , breast cancer , strain (injury) , pattern recognition (psychology) , medicine , cancer , biology , optics , physics , genetics
One fourth of the cancer detected in women worldwide is breast cancer which leads this as a major threat for women. There are many methods of detecting cancer among which ultra-sound strain imaging is one of the promising techniques. However, in strain sequence, not all the frames show clear tumor visibility. Consequently, in this paper we tested some well-defined algorithms to find only those frames where the tumor is comparatively clearly visible. We have used Mean Pixel Difference (MPD) and Gray- Level Co-occurrence Matrix (GLCM) to find the frames with better tumor visibility. We have tested our methods in several real-life cases and the results have been examined by a professional doctor. The MPD has an accuracy of 96.2% and the GLCM. Contrast has that of 55.55%.
GUB JOURNAL OF SCIENCE AND ENGINEERING, Vol 7, Dec 2020 P 8-13