Magnetic Resonance Imaging Breast Scan Classification based on Texture Features and Long Short-Term Memory Model
علي مجيد حسن
Authors : Assist. Prof. Ali Majeed Hasan
The aim of study is building new program for processing MRI images using MATLAB and to investigate different breast MRI detection algorithms that inform normal and abnormal scans of MRI. In this research an algorithm is proposed to extract texture feature and inform normal and abnormal scans of MRI. First, the MRI scans are pre- processed by image enhancement, intensity normalization, background segmentation and detection of mirror symmetry of breast. Second, the proposed gray level co- occurrence matrix (GLCM) and gray level run length matrix (GLRLM) methods are used to extract texture features from MRI T2-weighted and STIR images. Finally, these features are classified into normal and abnormal by using long short term memory (LSTM) model. The research will be validated using 326 datasets that downloaded from cancer imaging archive (TCIA). The achieved classification accuracy was 98.80%. Key Words: Deep Learning, GLCM, GLRLM, MRI Breast Scans, Feature Extraction, LSTM Classifier.

(FULL ARTICLE LINK) Read more ...
July 2021