AURORA: 自适应统一表示用于鲁棒超声分析

Ultrasound images exhibit significant variations across scanners, operators, and anatomical targets, often leading to poor generalization of models trained in one setting to new hospitals and clinical conditions. The Foundation Model Challenge for Ultrasound Image Analysis (FMC-UIA) highlights this challenge by requiring a single model to perform multiple tasks, including segmentation, detection, classification, and landmark regression across diverse organs and datasets. A unified multi-task framework is proposed, based on a transformer visual encoder from the Qwen3-VL family. Intermediate token features are projected into spatial feature maps and fused using a lightweight multi-scale feature pyramid, enabling both pixel-level predictions and global reasoning. This framework aims to enhance the robustness and adaptability of ultrasound image analysis.