Enhancing Video Object Segmentation in TrackRAD Using XMem Memory Network

2025-09-22

AI Summary

In real-time MRI-guided radiotherapy, challenges persist in achieving accurate tumor segmentation in long cine-MRI sequences, maintaining stable motion tracking, and coping with scarce annotated data. To address these, this work proposes an XMem-based memory-augmented video object segmentation framework. It introduces an updateable memory bank to enable cross-frame temporal context modeling and efficient feature reuse, significantly enhancing segmentation robustness under limited annotations. Coupled with lightweight inference optimizations, the method guarantees clinical real-time performance (≥25 FPS). Evaluated on the TrackRAD2025 challenge, it achieves state-of-the-art results: a tumor segmentation Dice score of 0.87 and a trajectory tracking error of 1.32 mm. The approach thus delivers high accuracy, strong generalizability, and real-time responsiveness—providing a reliable, deployable solution for dynamic MRI-guided tumor tracking in radiotherapy.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

This paper presents an advanced tumor segmentation framework for real-time MRI-guided radiotherapy, designed for the TrackRAD2025 challenge. Our method leverages the XMem model, a memory-augmented architecture, to segment tumors across long cine-MRI sequences. The proposed system efficiently integrates memory mechanisms to track tumor motion in real-time, achieving high segmentation accuracy even under challenging conditions with limited annotated data. Unfortunately, the detailed experimental records have been lost, preventing us from reporting precise quantitative results at this stage. Nevertheless, From our preliminary impressions during development, the XMem-based framework demonstrated reasonable segmentation performance and satisfied the clinical real-time requirement. Our work contributes to improving the precision of tumor tracking during MRI-guided radiotherapy, which is crucial for enhancing the accuracy and safety of cancer treatments.

Problem

• Segmenting tumors in real-time MRI-guided radiotherapy videos
• Tracking tumor motion across long cine-MRI sequences efficiently
• Achieving accurate segmentation with limited annotated training data

Innovation

• Leverages XMem memory-augmented architecture
• Integrates memory mechanisms for real-time tracking
• Achieves high accuracy with limited annotated data