Segmentation and Color ROI Extraction from Breast Imaging Datasets for Cancer Classification
Thành Tâm Nguyên, Thanh Hai Nguyen, Thanh Hai Nguyen, Thanh Hai Nguyen · Engineering Technology & Applied Science Research · 2025
Breast cancer is a serious health concern worldwide, particularly affecting many women. Therefore, using image processing techniques to enhance mammographic images is essential for accurate and early diagnosis. This study proposes a method for segmenting and extracting the Gray Region of Interest (GROI) from mammographic images and creating a corresponding Color Region of Interest (CROI) to improve classification performance. An EfficientNet-B7 model is used to classify image sets containing CROI. To accurately evaluate the effectiveness of CROI compared to GROI, the proposed method is applied to five categories of mammography image sets before training the EfficientNet-B7 model. Specifically, an automated algorithm is introduced to determine the thresholds values for extracting the GROI. The CROI is the generated using the proposed Identification and Comparison (IaC) algorithm. The results show that the classification accuracy improves from 84.3% with GROI to 92.6% with CROI, demonstrating the effectiveness of the proposed enhancement method for breast cancer image classification.