Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound

Automated breast ultrasound (ABUS) is being rapidly utilized for screening and diagnosing cancer. Breast masses, including cancers shown in ABUS scans, often appear as irregular hypoechoic areas that are hard to distinguish from background shadings. We propose a novel branch network architecture incorporating segmentation information of masses the training process. The integrated into neural network, providing spatial attention effect. boosts performance existing classifiers, helping learn meaningful features around target mass. For information, we leverage radiology reports without additional labeling efforts. reports, which generated medical image reading process, should include characteristics such shape orientation, template mask can be created rule-based manner. Experimental results show proposed with significantly improves classifiers. also provide qualitative interpretation method by visualizing effect on objects.