Texture Feature Extraction and Classification of Brain Neoplasm in MR Images using Machine Learning Techniques
Padmavathi K1, Maya V Karki2
1Padmavathi K*, Department. of EC, NMAMIT, Nitte, Karnataka, India.
2Maya V Karki, Department. of EC, Ramaiah Institute of Technology. Bengaluru.
Manuscript received on January 02, 2020. | Revised Manuscript received on January 15, 2020. | Manuscript published on January 30, 2020. | PP: 2319-2325 | Volume-8 Issue-5, January 2020. | Retrieval Number: E5704018520/2020©BEIESP | DOI: 10.35940/ijrte.E5704.018520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Identification of pathology in brain such as tumor lesions is a tedious task. MRI is most of the time chosen medical imaging procedure that often pacts with lenient tissues such as brain tissues, tendons and ligaments. This study aims at texture feature extraction and segregation of brain tumor cases into benign and malignant conditions. The stages involved are segmentation, feature extraction and classification. K-means clustering method is preferred for segmentation and selecting the required region of interest. The textural information is captured from region of interest using GLCM, HOG and LBP patterns. ANN, SVM and k-NN classifiers are used to analyze performance accuracy in classifying the tumor data into benign and malignant conditions in brain MR images. ANN with LM training algorithm provides high accuracy with best performance compared to other two classifiers in identifying benign and malignant conditions of tumors by using a combination of GLCM, LBP and HOG feature extraction process successfully. The recommended method is compared with few current approaches in terms of feature extraction and classification.
Keywords: Segmentation, Feature extraction, Tumor, MRI, classification.
Scope of the Article: Classification.