Fig. 9
From: A spine segmentation method based on scene aware fusion network
SAFNet has the ability to enhance the differentiation of each vertebra and intervertebral disc. SAFNet is designed to extract and learn semantic features from the input medical images, which can effectively distinguish different anatomical structures within the spine. Each row in the figure represents the middle sagittal slice of a subject. (SAFNet denotes scene aware fusion network; BG denotes background; MRI denote magnetic resonance imaging)