Multi-Channel Volumetric Neural Network for Knee Cartilage Segmentation in Cone-beam CT

TL;DR

This paper introduces a 3D convolutional neural network architecture for automatic knee cartilage segmentation in contrast-enhanced CT scans, aiming to facilitate osteoarthritis research by reducing manual effort.

Contribution

It presents a novel V-Net based model with a Tversky loss function tailored for cartilage segmentation in CT, addressing class imbalance and improving segmentation efficiency.

Findings

Achieved an average recall of 0.69 in cartilage segmentation

Demonstrated feasibility of CNN-based segmentation in contrast-enhanced CT

Reduced false positives by extracting largest connected components

Abstract

Analyzing knee cartilage thickness and strain under load can help to further the understanding of the effects of diseases like Osteoarthritis. A precise segmentation of the cartilage is a necessary prerequisite for this analysis. This segmentation task has mainly been addressed in Magnetic Resonance Imaging, and was rarely investigated on contrast-enhanced Computed Tomography, where contrast agent visualizes the border between femoral and tibial cartilage. To overcome the main drawback of manual segmentation, namely its high time investment, we propose to use a 3D Convolutional Neural Network for this task. The presented architecture consists of a V-Net with SeLu activation, and a Tversky loss function. Due to the high imbalance between very few cartilage pixels and many background pixels, a high false positive rate is to be expected. To reduce this rate, the two largest segmented point clouds are extracted using a connected component analysis, since they most likely represent the medial and lateral tibial cartilage surfaces. The resulting segmentations are compared to manual segmentations, and achieve on average a recall of 0.69, which confirms the feasibility of this approach.