Traditionally, video content analysis (https://en.wikipedia.org/wiki/Video_content_analysis) is implemented on specialized hardware which is capable of perfoming computationally demanding process of analysing video using computer vision algorithms. However, the recent advancements in processor designs (ARM, Intel, AMD, Nvidia) can be better exploited to perform real-time video analysis that is better targeted against the video surveillance domain. The end result will enable solutions to be affordable and better ROI for users.
Apart from generating realtime notifications about events in video ingested from security cameras, the side data available from video analysis will greatly enhance the chances of enabling users or ogranizations to increase profitability with business intelligence.
Eg: 1. How many shoppers were interested in the promotional product near the south east entrance of my store
2. What is attendance at my venue during the festive season in these dates
The idea is to build a scalable solution that covers both home (1-4 cameras) and medium sized (15-30 cameras) installations. Prototype development is in progress adhering to latest technologies both at the video analysis core logic and the presentation layers (UI, API etc.,).