Improving Image Restoration by Revisiting Global Information Aggregation
This addresses a critical but overlooked issue in image restoration models, offering a simple, cost-effective fix that enhances performance across multiple tasks without retraining.
The paper tackles the train-test inconsistency in global information aggregation for image restoration, where patch-based training and full-image inference cause performance drops, and proposes Test-time Local Converter (TLC) to convert global operations to local ones during inference, improving state-of-the-art results such as increasing PSNR from 32.92 dB to 33.57 dB on the GoPro dataset for single image deblurring.
Global operations, such as global average pooling, are widely used in top-performance image restorers. They aggregate global information from input features along entire spatial dimensions but behave differently during training and inference in image restoration tasks: they are based on different regions, namely the cropped patches (from images) and the full-resolution images. This paper revisits global information aggregation and finds that the image-based features during inference have a different distribution than the patch-based features during training. This train-test inconsistency negatively impacts the performance of models, which is severely overlooked by previous works. To reduce the inconsistency and improve test-time performance, we propose a simple method called Test-time Local Converter (TLC). Our TLC converts global operations to local ones only during inference so that they aggregate features within local spatial regions rather than the entire large images. The proposed method can be applied to various global modules (e.g., normalization, channel and spatial attention) with negligible costs. Without the need for any fine-tuning, TLC improves state-of-the-art results on several image restoration tasks, including single-image motion deblurring, video deblurring, defocus deblurring, and image denoising. In particular, with TLC, our Restormer-Local improves the state-of-the-art result in single image deblurring from 32.92 dB to 33.57 dB on GoPro dataset. The code is available at https://github.com/megvii-research/tlc.