Google Mediapipe public TFLite models implemented using Tensorflow (https://mediapipe.dev)
-
Hand Tracking
-
Related links
-
Hand Landmark 2D,3D
- Model Architecture (OK)
- Set Pretrained weights (OK)
- Convert TFLite to frozen graph (OK)
-
Palm(Hand) Detection
- Model Architecture (OK)
- Set Pretrained weights (OK)
- Convert TFLite to frozen graph (OK)
- Create TFLite model without custom operations (OK)
-
-
Face Detection
-
Related links
-
Face Detection
- Model Architecture (OK)
- Set Pretrained weights (OK)
- Convert TFLite to frozen graph (OK)
-
-
Hair Segmentation
-
Related links
-
Hair Segmentation
- Model Architecture (TODO)
- Define custom layers "MaxUnpooling2D" & "MaxPoolingWithArgmax2D"
- Set Pretrained weights (TODO)
- Convert TFLite to frozen graph(pb file) (TODO)
- Create TFLite model without custom operations (TODO)
- Model Architecture (TODO)
-