Deep Learning for Diagnosing Developmental Dysplasia of the Hip
To resolve the temporal diversity and pathological deformity challenge in X-ray based diagnosis of DDH, convert the detection of a landmark to the detection/segmentation of the landmark’s local neighborhood patch. Spatial attention and Adversarial learning are also employed to deal with the false positive detection.
More InfoDeep Learning for Pediatric Bone Age Assessment
Propose a novel weakly-supervised learning approach for Bone Age Assessment, which can discover and extract the discriminative bone parts without human prior. State-of-art performance has been achieved, and the proposed method exhibits interpretability, consistency and inspiration with human prior knowledge.
More Info