mars

MARS

Input

(Video from HMDB51 : https://serre-lab.clps.brown.edu/resource/hmdb-a-large-human-motion-database/)

Shape : (1, 3, duration, 112, 112))

Output

class_count=3
+ idx=0
  category=15[golf ]
  prob=6.96644401550293
+ idx=1
  category=43[swing_baseball ]
  prob=3.26068377494812
+ idx=2
  category=2[catch ]
  prob=2.2060546875

The accuracy of each category. (HMDB51)

Shape : (1, 51)

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 mars.py

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 mars.py --video VIDEO_PATH

Reference

MARS: Motion-Augmented RGB Stream for Action Recognition

HMDB: a large human motion database

Framework

Pytorch

Model Format

ONNX opset=10

Netron

mars.onnx.prototxt