Our new model ZOO works with DepthAI V3. Find out more in our documentation.
Model Details
Model Description
The MediaPipe Hand landmark model performs precise keypoint localization of 21 3D hand-knuckle coordinates inside the detected hand regions via regression, that is direct coordinate prediction. The model learns a consistent internal hand pose representation and is robust even to partially visible hands and self-occlusions.
Developed by: Google
Shared by:
Model type: Computer Vision
License:
Resources for more information:
Training Details
Training Data
For training the hand landmarker 3 different datasets were used:
In-the-wild dataset contains 6K images of a large variety, e.g. geographical diversity, various lighting conditions, and hand appearance. The limitation of this dataset is that it doesn’t contain complex articulation of hands.
In-house collected gesture dataset: This dataset contains 10K images that cover various angles of all physically possible hand gestures. The limitation of this dataset is that it’s collected from only 30 people with limited variation in background. The in-the-wild and in-house datasets are great complements to each other to improve robustness.
Synthetic dataset: To even better cover the possible hand poses and provide additional supervision for depth, we render a high-quality synthetic hand model over various backgrounds and map it to the corresponding 3D coordinates.
Testing Details
Metrics
Evaluation results are taken from .
Metric
Value
Mean Squared Error
11.83
Technical Specifications
Input/Output Details
Input:
Name: input_1
Info: NCHW BGR un-normalized image
Output:
Name: multiple (see NN archive)
Info: Hand keypoints, handedness (left or right hand), confidence score
Model Architecture
The model has a shared feature extractor and 3 separate heads for 3 outputs (hand landmarks, hand presence, and handedness). Each head is trained by correspondent datasets.
Please consult for more information on model architecture.
* Benchmarked with , using 2 threads (and the DSP runtime in balanced mode for RVC4).
* Parameters and FLOPs are obtained from the package.
Quantization
RVC4 version of the model was quantized using a custom dataset.
This was created by taking 50-image subset of dataset.
Utilization
Models converted for RVC Platforms can be used for inference on OAK devices.
DepthAI pipelines are used to define the information flow linking the device, inference model, and the output parser (as defined in model head(s)).
Below, we present the most crucial utilization steps for the particular model.
Please consult the docs for more information.
The example demonstrates how to build a 2-stage DepthAI pipeline consisting of a palm detection model and a hand landmark model.
It automatically downloads the models, creates a DepthAI pipeline, runs the inference, and displays the results using our DepthAI visualizer tool.
