arXiv:2512.11335v1 Announce Type: new 
Abstract: Ultrasound image segmentation is pivotal for clinical diagnosis, yet challenged by speckle noise and imaging artifacts. Recently, DINOv3 has shown remarkable promise in medical image segmentation with its powerful representation capabilities. However, DINOv3, pre-trained on natural images, lacks sensitivity to ultrasound-specific boundary degradation. To address this limitation, we propose FreqDINO, a frequency-guided segmentation framework that enhances boundary perception and structural consistency. Specifically, we devise a Multi-scale Frequency Extraction and Alignment (MFEA) strategy to separate low-frequency structures and multi-scale high-frequency boundary details, and align them via learnable attention. We also introduce a Frequency-Guided Boundary Refinement (FGBR) module that extracts boundary prototypes from high-frequency components and refines spatial features. Furthermore, we design a Multi-task Boundary-Guided Decoder (MBGD) to ensure spatial coherence between boundary and semantic predictions. Extensive experiments demonstrate that FreqDINO surpasses state-of-the-art methods with superior achieves remarkable generalization capability. The code is at https://github.com/MingLang-FD/FreqDINO.

تم تقديم FreqDINO كإطار عمل موجه بالتردد يهدف إلى تحسين تقسيم صور الموجات فوق الصوتية، وهو أمر أساسي للتشخيص السريري ولكنه غالبًا ما يعيقه الضوضاء والتشويش في الصور. تستخدم هذه الطريقة المبتكرة استراتيجية استخراج وتوافق التردد متعدد المقاييس لتعزيز إدراك الحدود والاتساق الهيكلي في صور الموجات فوق الصوتية.

FreqDINO se ha presentado como un marco de segmentación guiado por frecuencia destinado a mejorar la segmentación de imágenes de ultrasonido, que es esencial para el diagnóstico clínico pero a menudo obstaculizada por el ruido de speckle y los artefactos de imagen. Este enfoque innovador utiliza una estrategia de extracción y alineación de frecuencia a múltiples escalas para mejorar la percepción de los bordes y la consistencia estructural en las imágenes de ultrasonido.

FreqDINO a été introduit comme un cadre de segmentation guidé par la fréquence visant à améliorer la segmentation des images ultrasonores, essentielle pour le diagnostic clinique mais souvent entravée par le bruit de speckle et les artefacts d'imagerie. Cette approche innovante utilise une stratégie d'extraction et d'alignement de fréquence multi-échelle pour améliorer la perception des contours et la cohérence structurelle dans les images ultrasonores.

FreqDINO has been introduced as a frequency-guided segmentation framework aimed at improving ultrasound image segmentation, which is essential for clinical diagnosis but often hindered by speckle noise and imaging artifacts. This innovative approach utilizes a Multi-scale Frequency Extraction and Alignment strategy to enhance boundary perception and structural consistency in ultrasound images.

FreqDINO: Frequency-Guided Adaptation for Generalized Boundary-Aware Ultrasound Image Segmentation

Was this article worth reading? Share it

LucidQuery AI

Attentive AI

Deptho.ai

Dynamiq

Unifab

Blunge

Ready to build your own newsroom?