Code for the paper "Making Third Person Techniques Recognize First-Person Actions in Egocentric Videos" (ICIP 2018)

This is the github repository containing the code for the paper "Making Third Person Techniques Recognize First-Person Actions in Egocentric Videos" by Sagar Verma, Pravin Nagar, Divam Gupta, and Chetan Arora.

Requirements

The code has been tested on:

• Nvidia P5000 GPU
• Ubuntu 16.04 LTS
• Pytorch v0.4.0
• Opencv 3.0

Poster

Preprocessed dataset

1. GTEA
2. EGTEA Gaze Plus
3. Kitchen
4. UTE
5. ADL
6. HUJI

Publication

S. Verma, P. Nagar, D. Gupta, and C. Arora, "Making third person techniques recognize first-person actions in egocentric videos." International Conference on Image Processing (ICIP) PDF Poster

Citation

Please cite the following paper if you find this repository useful.

@article{Sagar2018ego,
  author      = {Sagar Verma and
                 Pravin Nagar and
                 Divam Gupta and
                 Chetan Arora},
  title       = {Making Third Person Techniques Recognize First-Person Actions in Egocentric Videos},
  booktitle   = {ICIP},
  year        = {2018}
}

Contact

For any queries, please contact

Sagar Verma: sagar15056@iiitd.ac.in