This repository is the zoo of image processing webworkers for javascript. You can use these workers as npm package.
Note. some module is not provided as webworker for safari because of it's restriction.
- Webworkers
- bodypix
- facemesh
- asciiart
- opencv
- PoseNet
- HandPose
- White-Box-Cartoon
- BiseNetv2 Celeb A Mask
- U^2-Net Portrait Drawing
- MODNet
- Google meet person segmentation
- Multi Barcode Scanner
- Super Resolution
- Blazeface
- MediaPipe Hands
- MediaPipe Face landmark detection
- MediaPipe BlazePose
- MediaPipe Mix
- MediaPipe Mix2
- TFLite Wasm
- Libs
- Experiments
- Reference
out2.mp4
Note: very heavy process. It will take 40second or more to process one frame. be patient... demo demo(slow)
White-box CartoonGAN
Copyright (C) Xinrui Wang All rights reserved. Licensed under the CC BY-NC-SA 4.0
license (https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode).
Commercial application is prohibited, please remain this license if you clone this repo
Note: Very heavy processing to open.
This project (code, pre-trained models, demos, etc.) is released under the Creative Commons Attribution NonCommercial ShareAlike 4.0 license.
NOTE: The license will be changed to allow commercial use after this work is accepted by a conference or a journal.
all all(slow) 96x160 96x160(slow) 128x128 128x128(slow) 144x256 144x256(slow) 256x256 256x256(slow)
-
Pixel3 https://youtu.be/IrPLMUuWaJk
-
Pixel4 https://youtu.be/Xxz1hFUAnKk
This project (code, pre-trained models, demos, etc.) is released under the Creative Commons Attribution NonCommercial ShareAlike 4.0 license.
all all(slow) lite lite(slow) full full(slow) mediapipe mediapipe(slow) tfjs tfjs(slow)
all all(slow) short short(slow) short_with_attention short_with_attention(slow) full full(slow) full_with_attention full_with_attention(slow) mediapipe mediapipe(slow) tfjs tfjs(slow)
all all(slow) lite lite(slow) full full(slow) heavy heavy(slow) mediapipe mediapipe(slow) tfjs tfjs(slow)
Compose Mediapipe models(hand, face, pose). This module bundles whole models in one file. If you want to split them, use MediaPipe Mix2.
Compose Mediapipe models(hand, face, pose) version2. This module load models from external path. This requires some of skills. If you want to use simply, use MediaPipe Mix.
Added Selfie Segmentationmodelcard
White-box CartoonGAN
Copyright (C) Xinrui Wang All rights reserved. Licensed under the CC BY-NC-SA 4.0
license (https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode).
Commercial application is prohibited, please remain this license if you clone this repo
https://flect-lab-web.s3-us-west-2.amazonaws.com/P01_wokers/tfl004_super_resolution/index.html
Faceswap by using facemesh worker
https://flect-lab-web.s3-us-west-2.amazonaws.com/P01_wokers/lib001_faceswap/index.html
With webworker, we can use models which based on the differenct tfjs models.
With webworker, we can improve performance when we use multiple models.
- Same as Exp.1
This repository was inspired by this site.
https://github.com/terryky/tfjs_webgl_app
Demo images are from pakutaso