Few-shot learning with deformable convolution for multi-scale lesion detection in mammography
MEDICAL PHYSICS, pp. 2970-2985, 2020.
Purpose Image-based breast lesion detection is a powerful clinical diagnosis technology. In recent years, deep learning architectures have achieved considerable success in medical image analysis however, they always require large-scale samples. In mammography images, breast lesions are inconspicuous, multiscale, and have blurred edges. Mo...More
Full Text (Upload PDF)
PPT (Upload PPT)