Accelerating Image Super-Resolution Networks with Pixel-Level Classification (ECCV 2024)

Abstract: In recent times, the need for effective super-resolution (SR) techniques has surged, especially for large-scale images ranging 2K to 8K resolutions. For DNN-based SISR, decomposing images into overlapping patches is typically necessary due to computational constraints. In such patch-decomposing scheme, one can allocate computational resources differently based on each patch's difficulty to further improve efficiency while maintaining SR performance. However, this approach has a limitation: computational resources is uniformly allocated within a patch, leading to lower efficiency when the patch contain pixels with varying levels of restoration difficulty. To address the issue, we propose the Pixel-level Classifier for Single Image Super-Resolution (PCSR), a novel method designed to distribute computational resources adaptively at the pixel level. A PCSR model comprises a backbone, a pixel-level classifier, and a set of pixel-level upsamplers with varying capacities. The pixel-level classifier assigns each pixel to an appropriate upsampler based on its restoration difficulty, thereby optimizing computational resource usage. Our method allows for performance and computational cost balance during inference without re-training. Our experiments demonstrate PCSR's advantage over existing patch-distributing methods in PSNR-FLOP trade-offs across different backbone models and benchmarks.

Dependencies

Python 3.7
Pytorch 1.13
NVIDIA GPU + CUDA
Python packages: pip install numpy opencv-python pandas tqdm fast_pytorch_kmeans

Demo

We provide a demo that performs 4x super-resolution on a single input LR image (supports only 4x scaling and .png format).

git clone https://huggingface.co/3587jjh/pcsr_carn

and then, specify your LR image path and output path like:

python demo.py --lr_path comic.png --output_path results

Output: flops: 66.3G (63.6 %) | max_flops: 104.1G (100 %)

For detailed demo and information, please refer to our GitHub page.

Citation

@inproceedings{jeong2025accelerating,
  title={Accelerating Image Super-Resolution Networks with Pixel-Level Classification},
  author={Jeong, Jinho and Kim, Jinwoo and Jo, Younghyun and Kim, Seon Joo},
  booktitle={European Conference on Computer Vision},
  pages={236--251},
  year={2025},
  organization={Springer}
}