Breast cancer characterization using region-based convolutional neural network with screening and diagnostic mammogram

• DOI: 10.12982/jams.2024.042
• Publication Date: 2024-06-11T09:15:28Z
• Authors: Jaroonroj Wongnil; Anchali Krisanach...; ; Rajalida Lipikorn

Background: Detection and classification of microcalcifications in breast tissues is crucial for early cancer diagnosis long-term treatment. Objective: This paper aims to propose a robust model capable detection calcifications digital mammogram images using Deep Convolutional Neural Networks (DCNN). Materials methods: An expert radiologist annotated the 3,265 clinical create comprehensive ground truth dataset comprising 2,500 annotations malignant benign calcifications. was utilized train our model, two-stage system incorporating Region-based Network (RCNN) with AlexNet support vector machines enhance system’s robustness. The proposed compared one-stage detection, utilizing YOLOv4 combined Cross-Stage Partial Darknet53 (CSPDarknet53) architecture. A separate 504 explicitly set aside testing. efficacy evaluated based on key performance metrics, including precision, recall, F1 score, mean average precision (mAP). Results: results showed that RCNN-2 could automatically identify categorize as or benign, outperforming models. RCNN-2’s overall effectiveness, by (mAP), achieved scores 0.82, 0.85, 0.83, 0.74, respectively. Conclusion: demonstrates very effective calcification images, especially high-dense images. YOLOv4, it can be concluded RCNN yields outstanding performance. helpful tool radiologists.