Simultaneous Bone and Shadow Segmentation Network using Task Correspondence Consistency

TL;DR

This paper introduces a novel end-to-end transformer-based network that simultaneously segments bone surfaces and their shadows in ultrasound images, leveraging task inter-dependencies for improved accuracy.

Contribution

It proposes a cross-task feature transfer block and a correspondence consistency loss to enhance joint segmentation of bones and shadows in ultrasound images.

Findings

Outperforms previous state-of-the-art in bone and shadow segmentation

Effectively leverages inter-task relationships for improved accuracy

Validated against expert annotations with superior results

Abstract

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.