COME: Dual Structure-Semantic Learning with Collaborative MoE for Universal Lesion Detection Across Heterogeneous Ultrasound Datasets

2025-08-13

AI Summary

To address severe inter-dataset interference, discriminative feature loss, and poor generalization under few-shot and cross-domain settings in multi-source heterogeneous ultrasound data, this paper proposes a Collaborative Mixture-of-Experts (MoE) framework. The framework features a dual-path expert architecture: a structure-semantic shared expert constructs a universal representation space, while source-specific experts preserve dataset-discriminative characteristics. Furthermore, it jointly models cross-dataset empirical distributions and learns universal priors to achieve joint structural and semantic optimization. Evaluated under three protocols—single-dataset, same-organ, and cross-organ—the framework achieves significant improvements in mean Average Precision (mAP) over state-of-the-art methods. Results demonstrate superior robustness to limited training samples and strong generalization capability on unseen domains and organs.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Conventional single-dataset training often fails with new data distributions, especially in ultrasound (US) image analysis due to limited data, acoustic shadows, and speckle noise. Therefore, constructing a universal framework for multi-heterogeneous US datasets is imperative. However, a key challenge arises: how to effectively mitigate inter-dataset interference while preserving dataset-specific discriminative features for robust downstream task? Previous approaches utilize either a single source-specific decoder or a domain adaptation strategy, but these methods experienced a decline in performance when applied to other domains. Considering this, we propose a Universal Collaborative Mixture of Heterogeneous Source-Specific Experts (COME). Specifically, COME establishes dual structure-semantic shared experts that create a universal representation space and then collaborate with source-specific experts to extract discriminative features through providing complementary features. This design enables robust generalization by leveraging cross-datasets experience distributions and providing universal US priors for small-batch or unseen data scenarios. Extensive experiments under three evaluation modes (single-dataset, intra-organ, and inter-organ integration datasets) demonstrate COME's superiority, achieving significant mean AP improvements over state-of-the-art methods. Our project is available at: https://universalcome.github.io/UniversalCOME/.

Problem

• Overcoming inter-dataset interference in universal ultrasound lesion detection
• Preserving dataset-specific features for robust downstream tasks
• Improving generalization across heterogeneous ultrasound datasets

Innovation

• Dual structure-semantic learning for universal representation
• Collaborative MoE with source-specific experts
• Cross-dataset experience for robust generalization