English

Efficient Perceptual Image Super Resolution: AIM 2025 Study and Benchmark

Computer Vision and Pattern Recognition 2025-10-15 v1

Abstract

This paper presents a comprehensive study and benchmark on Efficient Perceptual Super-Resolution (EPSR). While significant progress has been made in efficient PSNR-oriented super resolution, approaches focusing on perceptual quality metrics remain relatively inefficient. Motivated by this gap, we aim to replicate or improve the perceptual results of Real-ESRGAN while meeting strict efficiency constraints: a maximum of 5M parameters and 2000 GFLOPs, calculated for an input size of 960x540 pixels. The proposed solutions were evaluated on a novel dataset consisting of 500 test images of 4K resolution, each degraded using multiple degradation types, without providing the original high-quality counterparts. This design aims to reflect realistic deployment conditions and serves as a diverse and challenging benchmark. The top-performing approach manages to outperform Real-ESRGAN across all benchmark datasets, demonstrating the potential of efficient methods in the perceptual domain. This paper establishes the modern baselines for efficient perceptual super resolution.

Keywords

Cite

@article{arxiv.2510.12765,
  title  = {Efficient Perceptual Image Super Resolution: AIM 2025 Study and Benchmark},
  author = {Bruno Longarela and Marcos V. Conde and Alvaro Garcia and Radu Timofte},
  journal= {arXiv preprint arXiv:2510.12765},
  year   = {2025}
}

Comments

ICCV 2025 - AIM Workshop

R2 v1 2026-07-01T06:37:10.981Z