Spaces:
Running on Zero
Running on Zero
| title: iPhoneBlur Restore | |
| emoji: 📷 | |
| colorFrom: blue | |
| colorTo: gray | |
| sdk: gradio | |
| sdk_version: 6.30.0 | |
| app_file: app.py | |
| pinned: false | |
| license: apache-2.0 | |
| short_description: Sharpen motion-blurred phone photos at full resolution | |
| models: | |
| - Shafi99/iPhoneBlur_Finetune_Models | |
| datasets: | |
| - Shafi99/iPhoneBlur | |
| # iPhoneBlur Restore | |
| Upload a photo spoiled by handheld motion blur and drag the line to compare. | |
| The photo is restored at its native resolution and returned as PNG, so nothing is | |
| re-compressed after the model has done its work. Large photos are processed in | |
| overlapping 512 px tiles and blended. | |
| Trained on the [iPhoneBlur](https://huggingface.co/datasets/Shafi99/iPhoneBlur) | |
| benchmark: 7,400 blur and sharp pairs built from handheld phone video. Because the | |
| pairs come from one device, other cameras and other kinds of blur (out of focus, | |
| very dark scenes, heavy noise) may not improve. | |