SCUNet β€” Image Denoising (ONNX, full 8-variant bundle)

ONNX exports of SCUNet (Swin-Conv-UNet) β€” Kai Zhang et al., 2022. Hybrid CNN + Swin Transformer architecture for image denoising. This repo bundles all 8 published checkpoints from the upstream model_zoo/ so you get the full size / variant ladder in a single download.

Re-exported from upstream PyTorch weights. Provenance trail: Zhang et al. β†’ cszn/SCUNet model_zoo/*.pth β†’ torch.onnx.export (one pass per checkpoint) β†’ these files.

Toolchain: torch 2.4.x (CUDA 12.4), timm latest, einops latest, thop latest, onnx latest, onnxruntime>=1.17, opset 17, do_constant_folding=True. Full conversion script: scripts/export-kair.ps1 in the DatumIngest repo (runs once per .pth checkpoint via -Model scunet-color or -Model scunet-gray).

Credit: Kai Zhang, Yawei Li, Jingyun Liang, Jiezhang Cao, Yulun Zhang, Hao Tang, Deng-Ping Fan, Radu Timofte, Luc Van Gool. Paper: "Practical Blind Image Denoising via Swin-Conv-UNet and Data Synthesis", 2022.

What this repo contains

Each variant ships as an .onnx (small graph file) + .onnx.data (~70 MB of external tensor data) sibling pair β€” torch's ONNX exporter externalizes weights at the size/opset combination used here. Both files must be present in the same directory at load time β€” the .onnx references the .data by relative filename.

Color variants (5)

File pair Variant Training When to use
scunet_color_real_psnr.onnx[.data] Blind real-world, PSNR Mixed synthetic degradations (Gaussian + JPEG + downsampling), L1/L2 pixel loss Recommended default. General-purpose photo denoising. Stays faithful to input.
scunet_color_real_gan.onnx[.data] Blind real-world, GAN Same training data, adversarial + perceptual loss Consumer photo cleanup β€” sharper output, invents plausible texture. Skip when fidelity matters.
scunet_color_15.onnx[.data] Gaussian Οƒ=15 White Gaussian noise Οƒ=15 (light) Light noise (ISO grain). Beats blind on matched conditions; over-smooths cleaner inputs.
scunet_color_25.onnx[.data] Gaussian Οƒ=25 Οƒ=25 (moderate) Standard denoising-benchmark reference β€” apples-to-apples comparison with other papers at Οƒ=25.
scunet_color_50.onnx[.data] Gaussian Οƒ=50 Οƒ=50 (heavy) Extreme low-light / heavy-grain photos. Over-smooths anything cleaner.

Grayscale variants (3)

File pair Variant When to use
scunet_gray_15.onnx[.data] Gaussian Οƒ=15 Grayscale workflows (medical, document, B&W photo) at light noise.
scunet_gray_25.onnx[.data] Gaussian Οƒ=25 Standard grayscale-denoising benchmark level.
scunet_gray_50.onnx[.data] Gaussian Οƒ=50 Heavy-grain grayscale (astrophotography, degraded scans).

The grayscale variants are ~3Γ— cheaper to run than the color variants on grayscale inputs (they accept 1-channel input directly; the color variants need the gray channel replicated across RGB).

Input / output (all variants)

Color (in_nc=3) Gray (in_nc=1)
Input name image image
Input shape [batch, 3, H, W] (NCHW) [batch, 1, H, W]
Input dtype float32 float32
Input range [0, 1] RGB [0, 1] Y
Constraint H and W divisible by 8 H and W divisible by 8
Output name denoised denoised
Output shape [batch, 3, H, W] (same as input) [batch, 1, H, W]
Dynamic axes batch, height, width batch, height, width

All variants share the same forward-pass shape; the only differences are the input channel count and the trained weights.

How to use

import onnxruntime as ort
import numpy as np
from PIL import Image

# Pick a variant. Both the .onnx and .onnx.data must be present in
# the same directory β€” ORT resolves the external data automatically.
sess = ort.InferenceSession("scunet_color_real_psnr.onnx")

img = Image.open("noisy.jpg").convert("RGB")
W, H = img.size
W8, H8 = (W // 8) * 8, (H // 8) * 8                          # 8-align
img = img.crop((0, 0, W8, H8))

arr = np.asarray(img, dtype=np.float32) / 255.0              # HWC, [0,1]
arr = arr.transpose(2, 0, 1)[None, ...]                      # 1x3xHxW

result = sess.run(None, {"image": arr.astype(np.float32)})[0][0]
result = np.clip(result, 0.0, 1.0).transpose(1, 2, 0)
Image.fromarray((result * 255).astype(np.uint8)).save("denoised.jpg")

Which one should I use?

  • General-purpose photo denoising: scunet_color_real_psnr β€” blind, faithful, no guesswork required.
  • Consumer photo cleanup (subjectively pretty matters more than ground truth): scunet_color_real_gan.
  • Matched-Οƒ benchmark or known-noise scenario: pick the _15, _25, or _50 variant that matches your noise level.
  • Grayscale (medical / document / B&W): use scunet_gray_* directly β€” ~3Γ— faster than the color variant on gray inputs.
  • Comparison demos: the Οƒ-specialist variants are great for showing matched-vs-mismatched specialist behavior. Run the same noisy image through scunet_color_{15,25,50} and the differences are visually obvious.

For fixed-Οƒ Gaussian denoising in a research-benchmark context, SwinIR's swinir_denoising_color_25 is the apples-to-apples transformer counterpart. For denoising + sharpening as one step, look at NAFNet (opencv/deblurring_nafnet) β€” different task (deblur) but adjacent.

License

Apache-2.0 β€” same as the upstream cszn/SCUNet repo. LICENSE file included.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support