CrossMambaTuning: Synergistic Spatial and Cross-Layer Adaptation for Machine Vision Compression

2026-08-26

AI Summary

为解决预训练图像压缩模型适应下游任务时缺乏跨层协调的问题，提出CrossMambaTuning框架，通过状态空间模型和跨层交互机制实现高效微调。

Abstract

To reduce deployment cost and retraining overhead, adapting pretrained learned image compression (LIC) models to downstream machine vision tasks has attracted growing attention. However, existing methods typically insert fine-tuning modules independently into frozen backbones, lacking explicit mechanisms for cross-layer coordination. To address this limitation, we propose a novel framework named CrossMambaTuning, which integrates State Space Models with cross-layer interaction mechanisms for parameter-efficient fine-tuning. Specifically, we design an efficient Mamba adapter equipped with task-specific prompts and multi-scale branching to precisely capture both local features and global dependencies. Furthermore, we introduce a Scale-Invariant Cross-Layer Adapter (SICA) utilizing a parameter-sharing strategy to fuse task information across different scales and reduce redundancy. Extensive experiments demonstrate that CrossMambaTuning achieves state-of-the-art (SOTA) performance on multiple machine vision tasks, reducing parameter overhead by 72\% compared to SOTA methods. Code is available at https://github.com/rsr1123/CrossMambaTuning.

Problem

• pretrained learned image compression
• downstream machine vision tasks
• cross-layer coordination

Innovation

• State Space Models
• Cross-Layer Interaction
• Mamba Adapter
• Scale-Invariant Cross-Layer Adapter (SICA)
• Parameter-Efficient Fine-Tuning