Hugging Face's logo

qualcomm
/
MediaPipe-Pose-Estimation

Keypoint Detection
PyTorch
TF Lite
ONNX
real_time
android
Model card Files Community
1

MediaPipe-Pose-Estimation: Optimized for Mobile Deployment

Detect and track human body poses in real-time images and video streams

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

This model is an implementation of MediaPipe-Pose-Estimation found here.

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

Model Details

  • Model Type: Pose estimation
  • Model Stats:
    • Input resolution: 256x256
    • Number of parameters (MediaPipePoseDetector): 815K
    • Model size (MediaPipePoseDetector): 3.14 MB
    • Number of parameters (MediaPipePoseLandmarkDetector): 3.37M
    • Model size (MediaPipePoseLandmarkDetector): 12.9 MB
Model Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Precision Primary Compute Unit Target Model
MediaPipePoseDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 0.78 ms 0 - 121 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 0.841 ms 0 - 6 MB FP16 NPU MediaPipe-Pose-Estimation.so
MediaPipePoseDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 0.998 ms 0 - 34 MB FP16 NPU MediaPipe-Pose-Estimation.onnx
MediaPipePoseDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 0.558 ms 27 - 45 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 0.609 ms 0 - 20 MB FP16 NPU MediaPipe-Pose-Estimation.so
MediaPipePoseDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 0.738 ms 0 - 50 MB FP16 NPU MediaPipe-Pose-Estimation.onnx
MediaPipePoseDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 0.562 ms 0 - 17 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 0.497 ms 0 - 17 MB FP16 NPU Use Export Script
MediaPipePoseDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 0.754 ms 0 - 26 MB FP16 NPU MediaPipe-Pose-Estimation.onnx
MediaPipePoseDetector QCS8550 (Proxy) QCS8550 Proxy TFLITE 0.768 ms 0 - 5 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector QCS8550 (Proxy) QCS8550 Proxy QNN 0.821 ms 0 - 2 MB FP16 NPU Use Export Script
MediaPipePoseDetector SA7255P ADP SA7255P TFLITE 37.959 ms 0 - 17 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector SA7255P ADP SA7255P QNN 38.175 ms 0 - 10 MB FP16 NPU Use Export Script
MediaPipePoseDetector SA8255 (Proxy) SA8255P Proxy TFLITE 0.782 ms 0 - 7 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector SA8255 (Proxy) SA8255P Proxy QNN 0.823 ms 0 - 1 MB FP16 NPU Use Export Script
MediaPipePoseDetector SA8295P ADP SA8295P TFLITE 2.35 ms 0 - 10 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector SA8295P ADP SA8295P QNN 2.634 ms 0 - 5 MB FP16 NPU Use Export Script
MediaPipePoseDetector SA8650 (Proxy) SA8650P Proxy TFLITE 0.774 ms 0 - 12 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector SA8650 (Proxy) SA8650P Proxy QNN 0.82 ms 0 - 2 MB FP16 NPU Use Export Script
MediaPipePoseDetector SA8775P ADP SA8775P TFLITE 1.786 ms 0 - 17 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector SA8775P ADP SA8775P QNN 1.952 ms 0 - 6 MB FP16 NPU Use Export Script
MediaPipePoseDetector QCS8450 (Proxy) QCS8450 Proxy TFLITE 1.896 ms 0 - 20 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseDetector QCS8450 (Proxy) QCS8450 Proxy QNN 1.979 ms 0 - 14 MB FP16 NPU Use Export Script
MediaPipePoseDetector Snapdragon X Elite CRD Snapdragon® X Elite QNN 0.972 ms 0 - 0 MB FP16 NPU Use Export Script
MediaPipePoseDetector Snapdragon X Elite CRD Snapdragon® X Elite ONNX 1.042 ms 4 - 4 MB FP16 NPU MediaPipe-Pose-Estimation.onnx
MediaPipePoseLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 0.811 ms 0 - 8 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 0.907 ms 0 - 34 MB FP16 NPU MediaPipe-Pose-Estimation.so
MediaPipePoseLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 1.315 ms 0 - 9 MB FP16 NPU MediaPipe-Pose-Estimation.onnx
MediaPipePoseLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 0.603 ms 0 - 25 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 0.658 ms 0 - 25 MB FP16 NPU MediaPipe-Pose-Estimation.so
MediaPipePoseLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 0.973 ms 0 - 99 MB FP16 NPU MediaPipe-Pose-Estimation.onnx
MediaPipePoseLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 0.552 ms 0 - 23 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 0.61 ms 0 - 22 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 0.911 ms 0 - 45 MB FP16 NPU MediaPipe-Pose-Estimation.onnx
MediaPipePoseLandmarkDetector QCS8550 (Proxy) QCS8550 Proxy TFLITE 0.799 ms 0 - 10 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector QCS8550 (Proxy) QCS8550 Proxy QNN 0.911 ms 1 - 2 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector SA7255P ADP SA7255P TFLITE 17.242 ms 0 - 20 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector SA7255P ADP SA7255P QNN 17.628 ms 1 - 11 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector SA8255 (Proxy) SA8255P Proxy TFLITE 0.832 ms 0 - 190 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector SA8255 (Proxy) SA8255P Proxy QNN 0.947 ms 1 - 2 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector SA8295P ADP SA8295P TFLITE 1.5 ms 0 - 14 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector SA8295P ADP SA8295P QNN 1.839 ms 1 - 7 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector SA8650 (Proxy) SA8650P Proxy TFLITE 0.815 ms 0 - 45 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector SA8650 (Proxy) SA8650P Proxy QNN 0.936 ms 0 - 1 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector SA8775P ADP SA8775P TFLITE 1.638 ms 0 - 22 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector SA8775P ADP SA8775P QNN 1.938 ms 1 - 7 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector QCS8450 (Proxy) QCS8450 Proxy TFLITE 1.03 ms 0 - 22 MB FP16 NPU MediaPipe-Pose-Estimation.tflite
MediaPipePoseLandmarkDetector QCS8450 (Proxy) QCS8450 Proxy QNN 1.11 ms 1 - 21 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector Snapdragon X Elite CRD Snapdragon® X Elite QNN 1.108 ms 1 - 1 MB FP16 NPU Use Export Script
MediaPipePoseLandmarkDetector Snapdragon X Elite CRD Snapdragon® X Elite ONNX 1.352 ms 9 - 9 MB FP16 NPU MediaPipe-Pose-Estimation.onnx

Installation

This model can be installed as a Python 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_pose.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_pose.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_pose.export

Profiling Results
------------------------------------------------------------
MediaPipePoseDetector
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 0.8                    
Estimated peak memory usage (MB): [0, 121]               
Total # Ops                     : 106                    
Compute Unit(s)                 : NPU (106 ops)          

------------------------------------------------------------
MediaPipePoseLandmarkDetector
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 0.8                    
Estimated peak memory usage (MB): [0, 8]                 
Total # Ops                     : 219                    
Compute Unit(s)                 : NPU (219 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_pose import Model

# Load the model
model = Model.from_pretrained()
pose_detector_model = model.pose_detector
pose_landmark_detector_model = model.pose_landmark_detector

# Device
device = hub.Device("Samsung Galaxy S23")

# Trace model
pose_detector_input_shape = pose_detector_model.get_input_spec()
pose_detector_sample_inputs = pose_detector_model.sample_inputs()

traced_pose_detector_model = torch.jit.trace(pose_detector_model, [torch.tensor(data[0]) for _, data in pose_detector_sample_inputs.items()])

# Compile model on a specific device
pose_detector_compile_job = hub.submit_compile_job(
    model=traced_pose_detector_model ,
    device=device,
    input_specs=pose_detector_model.get_input_spec(),
)

# Get target model to run on-device
pose_detector_target_model = pose_detector_compile_job.get_target_model()
# Trace model
pose_landmark_detector_input_shape = pose_landmark_detector_model.get_input_spec()
pose_landmark_detector_sample_inputs = pose_landmark_detector_model.sample_inputs()

traced_pose_landmark_detector_model = torch.jit.trace(pose_landmark_detector_model, [torch.tensor(data[0]) for _, data in pose_landmark_detector_sample_inputs.items()])

# Compile model on a specific device
pose_landmark_detector_compile_job = hub.submit_compile_job(
    model=traced_pose_landmark_detector_model ,
    device=device,
    input_specs=pose_landmark_detector_model.get_input_spec(),
)

# Get target model to run on-device
pose_landmark_detector_target_model = pose_landmark_detector_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.

pose_detector_profile_job = hub.submit_profile_job(
    model=pose_detector_target_model,
    device=device,
)
pose_landmark_detector_profile_job = hub.submit_profile_job(
    model=pose_landmark_detector_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.

pose_detector_input_data = pose_detector_model.sample_inputs()
pose_detector_inference_job = hub.submit_inference_job(
    model=pose_detector_target_model,
    device=device,
    inputs=pose_detector_input_data,
)
pose_detector_inference_job.download_output_data()
pose_landmark_detector_input_data = pose_landmark_detector_model.sample_inputs()
pose_landmark_detector_inference_job = hub.submit_inference_job(
    model=pose_landmark_detector_target_model,
    device=device,
    inputs=pose_landmark_detector_input_data,
)
pose_landmark_detector_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-Pose-Estimation's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of MediaPipe-Pose-Estimation 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 API
Keypoint Detection
Inference API (serverless) does not yet support pytorch models for this pipeline type.