Three softwares for Video Analytics in Defense and Security Applications

Link to our presentation

A) SURAKSHAVYUH: ON-THE-EDGE PRODUCT FOR REAL TIME VIDEO SURVELLIANCE ANALYTICS

• This work on CCTV Video Analytics includes real-time intrusion detection, perimeter monitoring, loitering detection, object tracking, etc.

We have created an installer for our on-the-edge product, Suraksha Vyuha.

Download it using the following details

URL: https://www.cse.iitb.ac.in/~ganesh/surakshavyuh/compressed-sv-baseimage.tar.gz

Credentials to use the software can be provided upon acknowledgement/ request.

User manual:  SurakshaVyuha_UserManual.pdf

Installation guide:  SurakshaVyuha_Installation.pdf

For our offline summarization, search and analytics products (Jigyasa and Videosummy), the URLs (hosted on our server at IITB until you manage to get your own GPU server) are listed in B) and C) respectively:

B) JIGYASA:

1. For the video repository manager: http://videoanalytics.cse.iitb.ac.in/jigyasa/repo/login

• login & password: I can share upon acknowledgement/request from you or your team
• Here you can create a new case and subject it to offline analytics (on the server, the effect will be an annotated image like what you see in the attached file called
  " Screenshot from Jigyasa.png")

2. For the search and analysis utility: http://videoanalytics.cse.iitb.ac.in/jigyasa/search/login

• login & password: I can share upon acknowledgement/request from you or your team
• Here you can search for any of the following 12 queries " Woman, Person, Human Face, Man, vehicle, car, bus. truck, bicycle, motorcycle, boy, girl" (we can expand the scope of search queries later as required). You can also search for faces or entire person images.
• Next you can then navigate and "play" the search results (see example search navigation for bicycle  query image like what you see in the attached file called ScreenshotfromJigyasaCycleSearch.png)  

C) VIDEOSUMMY:

• Please try out the standalone software called VideoSummy  for summarizing long video footage and view it in seconds or minutes: .