arXiv:2606.24296v1 Announce Type: new 
Abstract: Cross-domain Few-shot Segmentation (CD-FSS) aims to learn generalizable segmentation capability from abundant annotated samples in the source domain, enabling accurate segmentation of novel classes in the target domain with only a few annotated samples. Existing CD-FSS methods mainly focus on mitigating feature distribution shifts caused by style gaps while ignoring significant differences in class semantic granularity and discriminative attributes across domains, leading to two key degradations in support-query matching: semantic over-alignment and attribute over-alignment. To this end, we propose the Dual Hierarchical Aggregation Network (DHANet), which comprises three key modules. First, the Hierarchical Spatial Aggregation (HSA) module performs multi-scale region aggregation of pixel features along the spatial dimension, generating hierarchical semantic-enhanced features to alleviate semantic over-alignment. Additionally, the HCA module conducts multi-scale attribute aggregation along the channel dimension, generating hierarchical attribute-enhanced features to mitigate attribute over-alignment. Finally, we propose the Online Probabilistic Semantic Bank (OPSB), which progressively constructs and updates class probability distributions from query predictions during inference, and samples multiple pseudo-prototypes as additional support information to mitigate insufficient support. Extensive experiments on four target-domain datasets demonstrate that our method achieves state-of-the-art performance.

تقدم دراسة جديدة شبكة التجميع الهرمي المزدوج (DHANet) التي تهدف إلى تحسين تقسيم الصور القليل من الأمثلة عبر المجالات (CD-FSS). تتناول هذه الطريقة التحديات المتعلقة بتغيرات توزيع الميزات والاختلافات في دقة المعاني بين المجالات، وتقترح وحدة التجميع المكاني الهرمي لتحسين مطابقة الدعم والاستعلام.

Un nuevo estudio presenta la Red de Agregación Jerárquica Dual (DHANet) destinada a mejorar la Segmentación por Pocos Ejemplos en Dominios Cruzados (CD-FSS). Este enfoque aborda los desafíos relacionados con los cambios en la distribución de características y las diferencias en la granularidad semántica entre dominios, proponiendo un módulo de Agregación Espacial Jerárquica para mejorar el emparejamiento soporte-consulta.

Une nouvelle étude présente le Dual Hierarchical Aggregation Network (DHANet) visant à améliorer la segmentation par peu d'exemples dans des domaines différents (CD-FSS). Cette approche traite les défis liés aux décalages de distribution des caractéristiques et aux différences de granularité sémantique entre les domaines, en proposant un module d'agrégation spatiale hiérarchique pour améliorer l'appariement support-requête.

A new study introduces the Dual Hierarchical Aggregation Network (DHANet) aimed at enhancing Cross-domain Few-shot Segmentation (CD-FSS). This approach addresses challenges in feature distribution shifts and semantic granularity differences across domains, proposing a Hierarchical Spatial Aggregation module to improve support-query matching.