deepView (sandbox demo v0.0.1)

-harnessing the preventive power of cctv networks

Project deepView uses localized CNN, Part Affinity Field based sekeleton generation and capsule CNN (ongoing) to classify & label real-time CCTV footage as 'Violent' or 'Non-Violent' - acting as an interface that can detect crime in real-time and alert law enforcement authorities.

Latest release:

Please check out our releases section to download the v0.0.1 release of the demo.

Full source:

Entire source is present in this repo.

Dependencies:

The following dependencies are necessary for the demo to work -

1. python3
2. scikit learn
3. pandas
4. pickle
5. g++

We also recommend downloading the full anaconda library as it contains all the above libraries.

System Requirements:

1. demo only works in windows os.
2. minimum 6 gb of free ram.
3. minimum nvidia gtx or higher (cuda enabled) graphics.

How To Run:

1) run models/getModels.bat 
2) run gui.py 

Troubleshoot Help:

Email: rishav_201500204@smit.smu.edu.in

References:

[1] Yun, Kiwon, et al. "Two-person interaction detection using body-pose features and multiple instance learning." Computer Vision and Pattern Recognition Workshops (CVPRW), 2012 IEEE Computer Society Conference on. IEEE, 2012.

[2]
@inproceedings{cao2017realtime,
  author = {Zhe Cao and Tomas Simon and Shih-En Wei and Yaser Sheikh},
  booktitle = {CVPR},
  title = {Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields},
  year = {2017}
}
[3]
@inproceedings{simon2017hand,
  author = {Tomas Simon and Hanbyul Joo and Iain Matthews and Yaser Sheikh},
  booktitle = {CVPR},
  title = {Hand Keypoint Detection in Single Images using Multiview Bootstrapping},
  year = {2017}
}
[4]
@inproceedings{wei2016cpm,
  author = {Shih-En Wei and Varun Ramakrishna and Takeo Kanade and Yaser Sheikh},
  booktitle = {CVPR},
  title = {Convolutional pose machines},
  year = {2016}
}