RAW-Adapter: Adapting Pre-trained Visual Model to Camera RAW Images and A Benchmark
This work addresses the challenge of leveraging RAW image data for computer vision tasks, offering a general framework that could benefit researchers and practitioners in fields like autonomous driving or photography, though it appears incremental by building on adapter-based methods.
The paper tackles the problem of adapting pre-trained visual models to camera RAW images by proposing RAW-Adapter, a framework that integrates learnable ISP modules and model-level adapters, and introduces RAW-Bench with 17 corruption types for evaluation, achieving robust performance across diverse scenarios as shown in extensive experiments.
In the computer vision community, the preference for pre-training visual models has largely shifted toward sRGB images due to their ease of acquisition and compact storage. However, camera RAW images preserve abundant physical details across diverse real-world scenarios. Despite this, most existing visual perception methods that utilize RAW data directly integrate image signal processing (ISP) stages with subsequent network modules, often overlooking potential synergies at the model level. Building on recent advances in adapter-based methodologies in both NLP and computer vision, we propose RAW-Adapter, a novel framework that incorporates learnable ISP modules as input-level adapters to adjust RAW inputs. At the same time, it employs model-level adapters to seamlessly bridge ISP processing with high-level downstream architectures. Moreover, RAW-Adapter serves as a general framework applicable to various computer vision frameworks. Furthermore, we introduce RAW-Bench, which incorporates 17 types of RAW-based common corruptions, including lightness degradations, weather effects, blurriness, camera imaging degradations, and variations in camera color response. Using this benchmark, we systematically compare the performance of RAW-Adapter with state-of-the-art (SOTA) ISP methods and other RAW-based high-level vision algorithms. Additionally, we propose a RAW-based data augmentation strategy to further enhance RAW-Adapter's performance and improve its out-of-domain (OOD) generalization ability. Extensive experiments substantiate the effectiveness and efficiency of RAW-Adapter, highlighting its robust performance across diverse scenarios.