efficientpose

EfficientPose

Input

(Image from https://github.com/daniegr/EfficientPose/blob/master/utils/MPII.jpg)

Model variant: RT Ailia input shape : (1, 224, 224, 3)
Range : [0, 1.0]

Output

• Confidence : (1, 224, 224, 16)
• Range : [0, 1.0]

Usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 efficientpose.py

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to change the name of the output file to save.

$ python3 efficientpose.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

If you want to specify the model variant, put the model variant after the --model_variant option.
You can only choose variants from 'rt','i','ii','iii','iv'.

$ python3 efficientpose.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH --model_variant rt

By adding the --video option, you can input the video.
If you pass 0 as an argument to VIDEO_PATH, you can use the webcam input instead of the video file.

$ python3 efficientpose.py --video VIDEO_PATH

Reference

Code repo for EfficientPose

Framework

Keras, TensorFlow, TensorFlow Lite or PyTorch

Model Format

ONNX opset = 10

Netron

EfficientPoseRT.onnx.prototxt

EfficientPoseI.onnx.prototxt

EfficientPoseII.onnx.prototxt

EfficientPoseIII.onnx.prototxt

EfficientPoseIV.onnx.prototxt