Reciprocal Landmark Detection and Tracking with Extremely Few Annotations

TL;DR

This paper introduces a novel end-to-end reciprocal detection and tracking model for echocardiography landmarks, effectively utilizing sparse annotations to improve localization accuracy and consistency in cardiac imaging.

Contribution

The paper presents a new reciprocal detection and tracking framework specifically designed for sparse labels in echocardiography, incorporating adversarial training to enhance performance.

Findings

Model outperforms existing methods in landmark localization accuracy.

Effective use of few annotated frames across entire cine sequences.

Reciprocal training improves detection and tracking consistency.

Abstract

Localization of anatomical landmarks to perform two-dimensional measurements in echocardiography is part of routine clinical workflow in cardiac disease diagnosis. Automatic localization of those landmarks is highly desirable to improve workflow and reduce interobserver variability. Training a machine learning framework to perform such localization is hindered given the sparse nature of gold standard labels; only few percent of cardiac cine series frames are normally manually labeled for clinical use. In this paper, we propose a new end-to-end reciprocal detection and tracking model that is specifically designed to handle the sparse nature of echocardiography labels. The model is trained using few annotated frames across the entire cardiac cine sequence to generate consistent detection and tracking of landmarks, and an adversarial training for the model is proposed to take advantage of these annotated frames. The superiority of the proposed reciprocal model is demonstrated using a series of experiments.