FVG-PT: Adaptive Foreground View-Guided Prompt Tuning for Vision-Language Models
2026-03-09 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors studied how a type of AI model called Vision-Language Models (VLMs) change their internal focus during a training method called prompt tuning. They found that problems happen when the model's attention to important parts of an image (the foreground) shifts too much. To fix this, they created a new method called Foreground View-Guided Prompt Tuning (FVG-PT), which helps the model keep better focus on key image parts and improves its predictions. Their tests showed this method works well across different models and datasets.
Vision-Language ModelsCLIPprompt tuningvisual attentionforegroundattention representationdistillationgeneralizationadaptation
Authors
Haoyang Li, Liang Wang, Siyu Zhou, Jiacheng Sun, Jing Jiang, Chao Wang, Guodong Long, Yan Peng
Abstract
CLIP-based prompt tuning enables pretrained Vision-Language Models (VLMs) to efficiently adapt to downstream tasks. Although existing studies have made significant progress, they pay limited attention to changes in the internal attention representations of VLMs during the tuning process. In this paper, we attribute the failure modes of prompt tuning predictions to shifts in foreground attention of the visual encoder, and propose Foreground View-Guided Prompt Tuning (FVG-PT), an adaptive plug-and-play foreground attention guidance module, to alleviate the shifts. Concretely, FVG-PT introduces a learnable Foreground Reliability Gate to automatically enhance the foreground view quality, applies a Foreground Distillation Compensation module to guide visual attention toward the foreground, and further introduces a Prior Calibration module to mitigate generalization degradation caused by excessive focus on the foreground. Experiments on multiple backbone models and datasets show the effectiveness and compatibility of FVG-PT. Codes are available at: https://github.com/JREion/FVG-PT