Skip to content

Latest commit

 

History

History
executable file
·
30 lines (17 loc) · 758 Bytes

README.md

File metadata and controls

executable file
·
30 lines (17 loc) · 758 Bytes
Raw

Activity Forecasting Python

This repository contains python implementation of activity forecasting proposed by Kitani et al. [1].

Execution environment

  • Language: Python (2.7.14)
  • Modules: Numpy, OpenCV

Training

Training (inverse optimal control)

python ioc.py

Inference (optimal control)

python oc.py

Optional: Convert XML file to Numpy array format.

If you want to use the original XML files of [1], please convert to numpy array format (*.npy). Copy to "convert_xml2npy.py" to the directory storing XML files, and execute the following:

python convert_xml2npy.py

Reference

  1. K. Kitani, et al., "Activity Forecasting," In Proc. of European Conference on Computer Vision (ECCV), pp. 201-214, 2012.