{
  "id": "2411.13548",
  "version": "v1",
  "published": "2024-11-20T18:56:24.000Z",
  "updated": "2024-11-20T18:56:24.000Z",
  "title": "HF-Diff: High-Frequency Perceptual Loss and Distribution Matching for One-Step Diffusion-Based Image Super-Resolution",
  "authors": [
    "Shoaib Meraj Sami",
    "Md Mahedi Hasan",
    "Jeremy Dawson",
    "Nasser Nasrabadi"
  ],
  "comment": "8 pages",
  "categories": [
    "cs.CV",
    "cs.LG"
  ],
  "abstract": "Although recent diffusion-based single-step super-resolution methods achieve better performance as compared to SinSR, they are computationally complex. To improve the performance of SinSR, we investigate preserving the high-frequency detail features during super-resolution (SR) because the downgraded images lack detailed information. For this purpose, we introduce a high-frequency perceptual loss by utilizing an invertible neural network (INN) pretrained on the ImageNet dataset. Different feature maps of pretrained INN produce different high-frequency aspects of an image. During the training phase, we impose to preserve the high-frequency features of super-resolved and ground truth (GT) images that improve the SR image quality during inference. Furthermore, we also utilize the Jenson-Shannon divergence between GT and SR images in the pretrained DINO-v2 embedding space to match their distribution. By introducing the $\\textbf{h}igh$- $\\textbf{f}requency$ preserving loss and distribution matching constraint in the single-step $\\textbf{diff}usion-based$ SR ($\\textbf{HF-Diff}$), we achieve a state-of-the-art CLIPIQA score in the benchmark RealSR, RealSet65, DIV2K-Val, and ImageNet datasets. Furthermore, the experimental results in several datasets demonstrate that our high-frequency perceptual loss yields better SR image quality than LPIPS and VGG-based perceptual losses. Our code will be released at https://github.com/shoaib-sami/HF-Diff.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2024-11-20T18:56:24.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "high-frequency perceptual loss",
      "one-step diffusion-based image super-resolution",
      "distribution matching",
      "loss yields better sr",
      "super-resolution methods achieve better"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 8,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}