Hugging Face's logo

qualcomm
/
MediaPipe-Hand-Detection

Object Detection
PyTorch
LiteRT
ONNX
real_time
android
Model card Files Community

MediaPipe-Hand-Detection: Optimized for Mobile Deployment

Real-time hand detection optimized for mobile and edge

The MediaPipe Hand Landmark Detector is a machine learning pipeline that predicts bounding boxes and pose skeletons of hands in an image.

This model is an implementation of MediaPipe-Hand-Detection found here.

This repository provides scripts to run MediaPipe-Hand-Detection on Qualcomm® devices. More details on model performance across various devices, can be found here.

Model Details

  • Model Type: Object detection
  • Model Stats:
    • Input resolution: 256x256
    • Number of parameters (MediaPipeHandDetector): 1.76M
    • Model size (MediaPipeHandDetector): 6.76 MB
    • Number of parameters (MediaPipeHandLandmarkDetector): 2.01M
    • Model size (MediaPipeHandLandmarkDetector): 7.71 MB
Model Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Precision Primary Compute Unit Target Model
HandDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 0.734 ms 0 - 31 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 0.718 ms 0 - 2 MB FP16 NPU MediaPipe-Hand-Detection.so
HandDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 1.003 ms 0 - 18 MB FP16 NPU MediaPipe-Hand-Detection.onnx
HandDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 0.54 ms 0 - 32 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 0.519 ms 0 - 19 MB FP16 NPU MediaPipe-Hand-Detection.so
HandDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 0.685 ms 0 - 39 MB FP16 NPU MediaPipe-Hand-Detection.onnx
HandDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 0.445 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 0.538 ms 1 - 20 MB FP16 NPU Use Export Script
HandDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 0.728 ms 1 - 25 MB FP16 NPU MediaPipe-Hand-Detection.onnx
HandDetector SA7255P ADP SA7255P TFLITE 24.662 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector SA7255P ADP SA7255P QNN 24.595 ms 1 - 10 MB FP16 NPU Use Export Script
HandDetector SA8255 (Proxy) SA8255P Proxy TFLITE 0.744 ms 0 - 30 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector SA8255 (Proxy) SA8255P Proxy QNN 0.712 ms 1 - 3 MB FP16 NPU Use Export Script
HandDetector SA8295P ADP SA8295P TFLITE 1.741 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector SA8295P ADP SA8295P QNN 1.677 ms 0 - 18 MB FP16 NPU Use Export Script
HandDetector SA8650 (Proxy) SA8650P Proxy TFLITE 0.738 ms 0 - 29 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector SA8650 (Proxy) SA8650P Proxy QNN 0.71 ms 1 - 4 MB FP16 NPU Use Export Script
HandDetector SA8775P ADP SA8775P TFLITE 1.575 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector SA8775P ADP SA8775P QNN 1.504 ms 1 - 11 MB FP16 NPU Use Export Script
HandDetector QCS8275 (Proxy) QCS8275 Proxy TFLITE 24.662 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector QCS8275 (Proxy) QCS8275 Proxy QNN 24.595 ms 1 - 10 MB FP16 NPU Use Export Script
HandDetector QCS8550 (Proxy) QCS8550 Proxy TFLITE 0.736 ms 0 - 31 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector QCS8550 (Proxy) QCS8550 Proxy QNN 0.705 ms 1 - 4 MB FP16 NPU Use Export Script
HandDetector QCS9075 (Proxy) QCS9075 Proxy TFLITE 1.575 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector QCS9075 (Proxy) QCS9075 Proxy QNN 1.504 ms 1 - 11 MB FP16 NPU Use Export Script
HandDetector QCS8450 (Proxy) QCS8450 Proxy TFLITE 1.429 ms 0 - 26 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandDetector QCS8450 (Proxy) QCS8450 Proxy QNN 1.41 ms 1 - 25 MB FP16 NPU Use Export Script
HandDetector Snapdragon X Elite CRD Snapdragon® X Elite QNN 0.869 ms 1 - 1 MB FP16 NPU Use Export Script
HandDetector Snapdragon X Elite CRD Snapdragon® X Elite ONNX 0.987 ms 4 - 4 MB FP16 NPU MediaPipe-Hand-Detection.onnx
HandLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 1.009 ms 0 - 64 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 1.006 ms 1 - 3 MB FP16 NPU MediaPipe-Hand-Detection.so
HandLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 1.339 ms 0 - 38 MB FP16 NPU MediaPipe-Hand-Detection.onnx
HandLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 0.757 ms 0 - 37 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 0.752 ms 1 - 20 MB FP16 NPU MediaPipe-Hand-Detection.so
HandLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 0.941 ms 0 - 33 MB FP16 NPU MediaPipe-Hand-Detection.onnx
HandLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 0.764 ms 0 - 27 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 0.706 ms 1 - 19 MB FP16 NPU Use Export Script
HandLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 0.795 ms 1 - 28 MB FP16 NPU MediaPipe-Hand-Detection.onnx
HandLandmarkDetector SA7255P ADP SA7255P TFLITE 35.432 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector SA7255P ADP SA7255P QNN 35.402 ms 1 - 10 MB FP16 NPU Use Export Script
HandLandmarkDetector SA8255 (Proxy) SA8255P Proxy TFLITE 1.015 ms 0 - 63 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector SA8255 (Proxy) SA8255P Proxy QNN 1.004 ms 1 - 3 MB FP16 NPU Use Export Script
HandLandmarkDetector SA8295P ADP SA8295P TFLITE 2.291 ms 0 - 23 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector SA8295P ADP SA8295P QNN 2.212 ms 0 - 18 MB FP16 NPU Use Export Script
HandLandmarkDetector SA8650 (Proxy) SA8650P Proxy TFLITE 0.998 ms 0 - 64 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector SA8650 (Proxy) SA8650P Proxy QNN 1.037 ms 1 - 3 MB FP16 NPU Use Export Script
HandLandmarkDetector SA8775P ADP SA8775P TFLITE 2.237 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector SA8775P ADP SA8775P QNN 2.164 ms 1 - 11 MB FP16 NPU Use Export Script
HandLandmarkDetector QCS8275 (Proxy) QCS8275 Proxy TFLITE 35.432 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector QCS8275 (Proxy) QCS8275 Proxy QNN 35.402 ms 1 - 10 MB FP16 NPU Use Export Script
HandLandmarkDetector QCS8550 (Proxy) QCS8550 Proxy TFLITE 1.008 ms 0 - 63 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector QCS8550 (Proxy) QCS8550 Proxy QNN 1.003 ms 1 - 4 MB FP16 NPU Use Export Script
HandLandmarkDetector QCS9075 (Proxy) QCS9075 Proxy TFLITE 2.237 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector QCS9075 (Proxy) QCS9075 Proxy QNN 2.164 ms 1 - 11 MB FP16 NPU Use Export Script
HandLandmarkDetector QCS8450 (Proxy) QCS8450 Proxy TFLITE 1.902 ms 0 - 36 MB FP16 NPU MediaPipe-Hand-Detection.tflite
HandLandmarkDetector QCS8450 (Proxy) QCS8450 Proxy QNN 1.908 ms 1 - 30 MB FP16 NPU Use Export Script
HandLandmarkDetector Snapdragon X Elite CRD Snapdragon® X Elite QNN 1.237 ms 1 - 1 MB FP16 NPU Use Export Script
HandLandmarkDetector Snapdragon X Elite CRD Snapdragon® X Elite ONNX 1.36 ms 6 - 6 MB FP16 NPU MediaPipe-Hand-Detection.onnx

Installation

Install the package via pip:

pip install qai-hub-models

Configure Qualcomm® AI Hub to run this model on a cloud-hosted device

Sign-in to Qualcomm® AI Hub with your Qualcomm® ID. Once signed in navigate to Account -> Settings -> API Token.

With this API token, you can configure your client to run models on the cloud hosted devices.

qai-hub configure --api_token API_TOKEN

Navigate to docs for more information.

Demo off target

The package contains a simple end-to-end demo that downloads pre-trained weights and runs this model on a sample input.

python -m qai_hub_models.models.mediapipe_hand.demo

The above demo runs a reference implementation of pre-processing, model inference, and post processing.

NOTE: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above).

%run -m qai_hub_models.models.mediapipe_hand.demo

Run model on a cloud-hosted device

In addition to the demo, you can also run the model on a cloud-hosted Qualcomm® device. This script does the following:

  • Performance check on-device on a cloud-hosted device
  • Downloads compiled assets that can be deployed on-device for Android.
  • Accuracy check between PyTorch and on-device outputs.
python -m qai_hub_models.models.mediapipe_hand.export

Profiling Results
------------------------------------------------------------
HandDetector
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 0.7                    
Estimated peak memory usage (MB): [0, 31]                
Total # Ops                     : 149                    
Compute Unit(s)                 : NPU (149 ops)          

------------------------------------------------------------
HandLandmarkDetector
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 1.0                    
Estimated peak memory usage (MB): [0, 64]                
Total # Ops                     : 158                    
Compute Unit(s)                 : NPU (158 ops)

How does this work?

This export script leverages Qualcomm® AI Hub to optimize, validate, and deploy this model on-device. Lets go through each step below in detail:

Step 1: Compile model for on-device deployment

To compile a PyTorch model for on-device deployment, we first trace the model in memory using the jit.trace and then call the submit_compile_job API.

import torch

import qai_hub as hub
from qai_hub_models.models.mediapipe_hand import Model

# Load the model
torch_model = Model.from_pretrained()

# Device
device = hub.Device("Samsung Galaxy S24")

# Trace model
input_shape = torch_model.get_input_spec()
sample_inputs = torch_model.sample_inputs()

pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])

# Compile model on a specific device
compile_job = hub.submit_compile_job(
    model=pt_model,
    device=device,
    input_specs=torch_model.get_input_spec(),
)

# Get target model to run on-device
target_model = compile_job.get_target_model()

Step 2: Performance profiling on cloud-hosted device

After compiling models from step 1. Models can be profiled model on-device using the target_model. Note that this scripts runs the model on a device automatically provisioned in the cloud. Once the job is submitted, you can navigate to a provided job URL to view a variety of on-device performance metrics.

profile_job = hub.submit_profile_job(
    model=target_model,
    device=device,
)

Step 3: Verify on-device accuracy

To verify the accuracy of the model on-device, you can run on-device inference on sample input data on the same cloud hosted device.

input_data = torch_model.sample_inputs()
inference_job = hub.submit_inference_job(
    model=target_model,
    device=device,
    inputs=input_data,
)
    on_device_output = inference_job.download_output_data()

With the output of the model, you can compute like PSNR, relative errors or spot check the output with expected output.

Note: This on-device profiling and inference requires access to Qualcomm® AI Hub. Sign up for access.

Deploying compiled model to Android

The models can be deployed using multiple runtimes:

  • TensorFlow Lite (.tflite export): This tutorial provides a guide to deploy the .tflite model in an Android application.

  • QNN (.so export ): This sample app provides instructions on how to use the .so shared library in an Android application.

View on Qualcomm® AI Hub

Get more details on MediaPipe-Hand-Detection's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of MediaPipe-Hand-Detection can be found here.
  • The license for the compiled assets for on-device deployment can be found here

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Object Detection
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support