Framework for YOLO–DETR–SAM Breast Cancer Detection and Segmentation Across Mammography
Journal of Al-Farabi for Engineering Sciences · 01.12.2025 · 4 (2) · 409-427
DOI: 10.59746/pmy0xn98 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Al-Farabi for Engineering Sciences, online erschienen am 01.12.2025. Verfasst von Rajaa Daami resen und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Breast-mass detection on mammograms remains challenging under breast-density variation and dataset shift. This manuscript proposes a four-phase framework integrating YOLOv8/YOLOv11, Deformable DETR/DINO-DETR, Weighted Boxes Fusion, and prompt-driven SAM/MedSAM segmentation with reciprocal external evaluation on VinDr-Mammo and CMMD. The original YOLO benchmark reported an mAP@50 of 0.793 for YOLOv11L and recall of 0.713 for YOLOv8X on a prior experimental split; these estimates require reassessment after patient-level repartitioning. Additional detector comparison, fusion, contour segmentation, cross-dataset transfer, and ablation analyses are specified as experimental extensions. Indicative numerical values presented for these extensions are provisional planning scenarios, not verified training outputs or patient-derived confidence intervals. The proposed workflow separates detection and segmentation metrics, requires expert reference contours for Dice/IoU analysis, and anticipates paired case-level uncertainty estimation. Its prospective methodological contribution is a controlled evaluation of complementary detector families and cross-dataset robustness; effectiveness cannot be established until the full framework is executed and its predictions are independently validated.
Zitieren
APA
resen, R. D., & Ibrahim, H. M. (2025). Framework for YOLO–DETR–SAM Breast Cancer Detection and Segmentation Across Mammography. Journal of Al-Farabi for Engineering Sciences, 4(2), 409-427. https://doi.org/10.59746/pmy0xn98
Vancouver
resen RD, Ibrahim HM. Framework for YOLO–DETR–SAM Breast Cancer Detection and Segmentation Across Mammography. Journal of Al-Farabi for Engineering Sciences. 2025;4(2):409-427. doi:10.59746/pmy0xn98
BibTeX
@article{resen2025,
author = {resen, Rajaa Daami and Ibrahim, Hassan muayad},
title = {Framework for YOLO–DETR–SAM Breast Cancer Detection and Segmentation Across Mammography},
journal = {Journal of Al-Farabi for Engineering Sciences},
year = {2025},
volume = {4},
number = {2},
pages = {409--427},
doi = {10.59746/pmy0xn98}
}
EUVIMED