Simultaneous Bone and Shadow Segmentation Network using Task Correspondence Consistency
This work addresses a domain-specific problem for ultrasound-guided orthopedic procedures, offering an incremental improvement by leveraging task interdependencies to enhance segmentation accuracy.
The paper tackles the challenge of segmenting bone surfaces and corresponding acoustic shadows in ultrasound images, which is difficult due to blurred responses and imaging artifacts, by proposing a single network with shared encoder and task-specific decoders that uses cross-task feature transfer and a consistency loss, resulting in outperforming previous state-of-the-art methods in validation against expert annotations.
Segmenting both bone surface and the corresponding acoustic shadow are fundamental tasks in ultrasound (US) guided orthopedic procedures. However, these tasks are challenging due to minimal and blurred bone surface response in US images, cross-machine discrepancy, imaging artifacts, and low signal-to-noise ratio. Notably, bone shadows are caused by a significant acoustic impedance mismatch between the soft tissue and bone surfaces. To leverage this mutual information between these highly related tasks, we propose a single end-to-end network with a shared transformer-based encoder and task independent decoders for simultaneous bone and shadow segmentation. To share complementary features, we propose a cross task feature transfer block which learns to transfer meaningful features from decoder of shadow segmentation to that of bone segmentation and vice-versa. We also introduce a correspondence consistency loss which makes sure that network utilizes the inter-dependency between the bone surface and its corresponding shadow to refine the segmentation. Validation against expert annotations shows that the method outperforms the previous state-of-the-art for both bone surface and shadow segmentation.