The Flow has not summarised this story yet — read it at arXiv Computer Vision.
arXiv:2608. 07749v1 Announce Type: cross Abstract: Parameter-efficient fine-tuning enables the adaptation of vision foundation models to biomedical tasks under limited computational resources, but a single low-rank update can constrain all task-specific changes to one narrow parameter subspace.
The paper introduces VPRef, the first cross‑domain benchmark for Referring Remote Sensing Image Segmentation, containing 46,972 language‑image‑annotation triplets with a three‑tier linguistic hierarchy. It proposes a parameter‑efficient adaptation method based on the Segment Anything Model and Low‑Rank Adaptation, using pseudo‑label self‑training for visual drift and random multi‑granularity prompt mixing for textual drift. Experiments show the approach improves cross‑domain segmentation while altering only 1.08 % of the base model’s parameters, offering a strong baseline for future research.
Rapid advancements in vision-language models have propelled Referring Remote Sensing Image Segmentation (RRSIS) to the forefront of Earth observation. However, practical deployments suffer severe perf...
arXiv:2607. 06918v1 Announce Type: cross Abstract: Pre-trained Vision Foundation Models (VFMs) provide strong visual representations for diverse downstream tasks.
The paper introduces FAN-LoRA, a Fourier‑adaptive nonlinear low‑rank adaptor designed to improve domain adaptation of vision foundation models like SAM for medical imaging. By decoupling frequency components into a low‑pass B‑spline branch for global structure and a high‑pass Fourier branch for local texture, FAN‑LoRA addresses performance drops caused by domain shifts. Experiments on three cross‑modality and cross‑center benchmarks show that FAN‑LoRA outperforms existing PEFT methods, achieving higher Dice scores and lower boundary errors while remaining computationally efficient.
The core challenge of heterogeneous change detection in remote sensing imagery lies in effectively decoupling genuine land-cover changes from significant modal disparities caused by distinct imaging mechanisms. These intrinsic inconsistencies are prone to introducing pseudo-changes, thereby constraining detection accuracy.