ClearView: Image Deraining with ResNet50-UNet (Mixed-Domain)
ClearView demo showcase: rainy input vs. derained output across four scenes (ResNet50-UNet)
Task
Domain
Params
License
A U-Net [8] with a ResNet-50 [9] encoder (ImageNet-pretrained, fine-tuned end to end), the largest of ClearView's three ResNet-backbone variants, trained on a blended synthetic + real-world rain set, selecting checkpoints against a blended real-world validation metric rather than a single benchmark, for a model that holds up across domains instead of maxing out one dataset's quirks. Not a mirror of someone else's checkpoint, see Citation.
~62% synthetic / ~38% real by effective sampling weight (before oversampling: ~77%/23% by raw pair count).
Checkpoint selection uses a separate blended validation set (recipe): SPA-Data val (capped to 150 of 1,000 pairs so it can't dominate), RealRain-1k-H/L validation (112 each), and Rain100L (100) as a synthetic sanity anchor.
Detailed Test-Set Metrics
Full per-dataset breakdown across all 6 tracked metrics, computed on each source's own held-out test/eval split (not the blended validation set used for checkpoint selection during training).
One negative Rain Removal Rate value here (SPA-Data: -0.476) is worth calling out explicitly: PSNR/SSIM are still decent there, but the Sobel-residual metric says this checkpoint added slightly more high-frequency error than it removed on that set specifically, likely mild over-sharpening/artifacting from the ResNet encoder path rather than a training failure. Unlike ResNet18/34-UNet, Rain100L's Rain Removal Rate here is actually positive (0.045), the smallest positive value in the whole ClearView lineup, but positive nonetheless. See Use Cases below.
Metric definitions:
PSNR (Peak Signal-to-Noise Ratio, dB): pixel-level fidelity between the restored output and ground truth. Higher is better. The standard image-restoration metric, but insensitive to structural/perceptual quality on its own.
SSIM (Structural Similarity Index, 0-1): perceptual similarity based on luminance, contrast, and structure. Higher is better, tracks human-perceived quality more closely than PSNR.
MAE (Mean Absolute Error, normalized [0,1] pixel space): average per-pixel absolute intensity difference. Lower is better, less sensitive to outlier pixels than MSE.
MSE (Mean Squared Error, normalized [0,1] pixel space): average per-pixel squared intensity difference, the term PSNR is derived from. Lower is better, penalizes large errors more heavily than MAE.
Rain Removal Rate: ClearView-specific metric. Compares the Sobel-gradient ("high-frequency") residual energy of the output vs. ground truth against that of the input vs. ground truth: 1 - (residual after) / (residual before). 1.0 means perfect rain removal, 0.0 means no change, negative means the model added more high-frequency error than it removed (e.g. hallucinated detail or over-sharpening). Higher is better.
NIQE (Natural Image Quality Evaluator): no-reference perceptual quality score. Its pristine reference statistics are refit per test set from that set's own clean images, so NIQE values are only comparable within the same row, not across rows, a NIQE of 800 on one dataset and 10 on another does not mean one output is 80x worse, the reference scale itself differs per dataset. Lower is better within a given row.
ClearView Model Comparison
How this model compares to the rest of the ClearView model zoo (PSNR / SSIM), same test sets and evaluation protocol for every model. Columns sorted highest to lowest by average PSNR across the 9 rain-only test sets.
All ClearView models trained under the identical mixed-domain recipe, only batch size/accumulation steps vary per architecture size. Histoformer is included as an external, inference-only reference point (original authors' checkpoint, not trained under this recipe), its strong AllWeather (rain+fog) score and comparatively weak scores on the rain-only test sets reflect that it was trained on a rain+fog+snow mix, while ClearView's current mix is rain-only. ResNet50-UNet is the largest of the three ClearView ResNet-backbone variants at 73.3M params, but that extra capacity doesn't translate into a clean win, it edges out ResNet34/18-UNet only slightly (30.63 vs. 30.45 vs. 30.23 avg PSNR across the 9 rain-only sets) and still trails NAFNet (Small) at 1.1M params. Likely a ResNet-stem effect (the stock ResNet stem's aggressive early downsampling, a stride-2 7x7 conv followed by a stride-2 max-pool, can blur or alias away the thin rain streaks it's supposed to detect before any residual block runs) shared across all three ResNet variants regardless of depth, rather than a capacity issue, still a working hypothesis rather than a confirmed cause.
Use Cases
Good for: dashcam/surveillance footage, photo restoration, real-world rain (not just synthetic-style streaks); a reference point for how a plain ImageNet-pretrained ResNet encoder compares against ClearView's purpose-built architectures on this task, at the largest capacity point in that comparison.
Limitations: no temporal consistency for video (processes frames independently); AllWeather rain+fog is an explicit out-of-scope stress test, fog was not in the training mix and this model is not expected to handle it; trails every other ClearView model on rain-only test sets despite being the largest by params (73.3M); negative Rain Removal Rate on SPA-Data, PSNR/SSIM are still reasonable there but the residual high-frequency error metric suggests mild over-sharpening/artifacting rather than clean streak removal on that set; largest and slowest of the three ResNet-backbone variants, needs a GPU with meaningfully more VRAM than ResNet18/34-UNet for a modest quality gain.
