Image and video analysis as a service framework-aware AI
Computer vision and image analysis does not easily transition into the commercial landscape because it is disconnected from the end users. It is difficult to produce models that account for varying scenes and to interpret events of interest. One example would be in video surveillance with data consuming networks. Today, surveillance cameras are deployed in large numbers and analytics are run on live streams for monitoring and decision making and on stored videos for forensic purposes. We propose a scalable implementation that allows video surveillance operators to run analytics on multiple video and live streams simultaneously with results of interest obtained directly by end users.
  • Video Security in retail and industrial settings
Problems Addressed
  • Current disconnect between end user and data acquisition system
Competitive Advantages
  • API web server acts as an interface between all the components and saves the videos in database and allows for Create, Read, Update, Delete (CRUD) operations.
  • Decouples the front end from the framework implementation and allows for the construction of an assortment of clients including web applications, desktop applications, and mobile applications.
  • Provisional Patent Application Filed
Case ID
2020- 003
Shishir Shah
Professor, Computer Science