# To-Do list for Human library ## Work in Progress
### Models #### Hand - Evaluate and switch default default model from `handdetect` to `handtrack` #### Body - Implement new variations of `BlazePose` models - Add virtual box frame caching to `MoveNet` #### Face - Reimplement `BlazeFace`, `FaceMesh`, `Iris` with new pipeline and frame caching
### Backends #### WebGL - Optimize shader packing for WebGL backend: #### WASM - Backend WASM incorrect handling of `int32` tensors #### WebGPU Implementation of WebGPU backend Experimental support only until support is officially added in Chromium - Evaluate WGSL vs GLSL for WebGPU - Backend WebGPU missing kernel ops - Backend WebGPU incompatible with web workers - Backend WebGPU incompatible with sync read calls
### Exploring - Optical Flow: - TFLite Models:


## Known Issues
### Face Detection Enhanced rotation correction for face detection is not working in NodeJS due to missing kernel op in TFJS Feature is automatically disabled in NodeJS without user impact - Backend NodeJS missing kernel op `RotateWithOffset` ### Hand Detection Enhanced rotation correction for hand detection is not working in NodeJS due to missing kernel op in TFJS Feature is automatically disabled in NodeJS without user impact - Backend NodeJS missing kernel op `RotateWithOffset` ### Body Detection MoveNet MultiPose model does not work with WASM backend due to missing F32 broadcast implementation - Backend WASM missing F32 broadcat implementation ### Object Detection Object detection using CenterNet or NanoDet models is not working when using WASM backend due to missing kernel ops in TFJS - Backend WASM missing kernel op `Mod` - Backend WASM missing kernel op `SparseToDense`