LGRNet: Local-Global Reciprocal Network for Uterine Fibroid Segmentation in Ultrasound Videos

2024-07-08

International Conference on Medical Image Computing and Computer-Assisted Intervention

AI Summary

To address low segmentation accuracy and insufficient temporal modeling in uterine fibroid segmentation from ultrasound videos, this paper proposes LGRNet—a local-global bidirectional interaction network. Methodologically, it introduces Cycle Neighborhood Propagation (CNP) to explicitly model local temporal dependencies and integrates Hilbert Selective Scanning (HilbertSS) for efficient long-range spatiotemporal context encoding; bottleneck query compression and path-aware attention are further incorporated to enhance computational efficiency and robustness. Contributions include: (1) the first benchmark dataset for ultrasound video-based fibroid segmentation, UFUV, comprising 100 expert-annotated video sequences; (2) state-of-the-art performance on UFUV and three public video panoptic segmentation (VPS) benchmarks, with average Dice score improvements of 2.3%–4.1%; and (3) full open-sourcing of code, pretrained models, and the UFUV dataset.

Technology Category

• Computer Vision: Segmentation
• Search and Optimization: Local Search
• Machine Learning: Multimodal Learning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Regular screening and early discovery of uterine fibroid are crucial for preventing potential malignant transformations and ensuring timely, life-saving interventions. To this end, we collect and annotate the first ultrasound video dataset with 100 videos for uterine fibroid segmentation (UFUV). We also present Local-Global Reciprocal Network (LGRNet) to efficiently and effectively propagate the long-term temporal context which is crucial to help distinguish between uninformative noisy surrounding tissues and target lesion regions. Specifically, the Cyclic Neighborhood Propagation (CNP) is introduced to propagate the inter-frame local temporal context in a cyclic manner. Moreover, to aggregate global temporal context, we first condense each frame into a set of frame bottleneck queries and devise Hilbert Selective Scan (HilbertSS) to both efficiently path connect each frame and preserve the locality bias. A distribute layer is then utilized to disseminate back the global context for reciprocal refinement. Extensive experiments on UFUV and three public Video Polyp Segmentation (VPS) datasets demonstrate consistent improvements compared to state-of-the-art segmentation methods, indicating the effectiveness and versatility of LGRNet. Code, checkpoints, and dataset are available at https://github.com/bio-mlhui/LGRNet

Problem

• Segmentation of uterine fibroids in ultrasound videos
• Propagation of long-term temporal context for lesion distinction
• Efficient aggregation of global and local temporal information

Innovation

• Local-Global Reciprocal Network for segmentation
• Cyclic Neighborhood Propagation for temporal context
• Hilbert Selective Scan for global context aggregation