Deep Adversarial Context-Aware Landmark Detection for Ultrasound Imaging

TL;DR

This paper introduces a deep adversarial multitask learning method for real-time prostate landmark detection in ultrasound images, improving accuracy and structural feasibility over previous approaches.

Contribution

The paper presents a novel adversarial multitask deep learning framework that jointly learns prostate landmarks and contours, enhancing robustness and accuracy in ultrasound imaging.

Findings

Achieved 92.6% Dice score with adversarial multitask approach.

Reduced mean distance error by 20% compared to landmark-only learning.

Operates in real-time on standard hardware.

Abstract

Real-time localization of prostate gland in trans-rectal ultrasound images is a key technology that is required to automate the ultrasound guided prostate biopsy procedures. In this paper, we propose a new deep learning based approach which is aimed at localizing several prostate landmarks efficiently and robustly. We propose a multitask learning approach primarily to make the overall algorithm more contextually aware. In this approach, we not only consider the explicit learning of landmark locations, but also build-in a mechanism to learn the contour of the prostate. This multitask learning is further coupled with an adversarial arm to promote the generation of feasible structures. We have trained this network using ~4000 labeled trans-rectal ultrasound images and tested on an independent set of images with ground truth landmark locations. We have achieved an overall Dice score of 92.6% for the adversarially trained multitask approach, which is significantly better than the Dice score of 88.3% obtained by only learning of landmark locations. The overall mean distance error using the adversarial multitask approach has also improved by 20% while reducing the standard deviation of the error compared to learning landmark locations only. In terms of computational complexity both approaches can process the images in real-time using standard computer with a standard CUDA enabled GPU.