An offline compression artifact remover and denoise application for Android
Models list available here, see examples for how to use them.
Caution
Running for an extended amount of time can make your device run hot, please monitor device temperature
Translate: https://translate.codeberg.org/projects/dejpeg
If you'd like to use models on PCs (Linux/Mac/Windows), look at chaiNNer (for FBCNN, install this custom node and use PyTorch models from the FBCNN repo).
This app will not be on the Play Store and I will not be involved with Developer "Verification", the code can be freely audited here.
There's an experimental version using the LiteRT backend with .tflite models. Conversion is difficult but the performance boost is significant for FBCNN and RMBG, SCUNet is incompatible and others have worse performance. You can download this version in Releases now merged with Oidn called extra (reason being APK size limitations with IzzyOnDroid)
For onnx-only builds (what's on IzzyOnDroid), run ./gradlew assembleOnnxOnlyDebug
To set up Intel® Oidn Denoising, see oidn/readme.md.
This is a GUI for 1x image processing models, used under their respective licenses
DeJPEG v4 and onward is licensed under AGPLv3-or-later Older versions (v3 and under) are under GPLv3-only.
Credits to @adrianerrea for a starting point, FBCNN and SCUNet creators plus all other model owners.
DeJPEG is not affiliated or related with Topaz DEJPEG or any other similarly named software/project. Although I've wondered if the term 'JPEG' is copyrighted/trademarked due to it literally being the acronym for Joint Photographic Experts Group, for this reason I might need to change the app's name if legal issues start to occur