![Two CCTV cameras](/media/003jniyu/cctv-1.jpg?center=0.69000125454773553,0.13333333333333333&mode=crop&width=378&height=330&format=jpg&quality=80&rnd=133077098497130000)
The Challenge
The world is generating massive amounts of video data, and mining it can be problematic. Automated search engines that rely on pre-tagged data, usually return too many low-quality matches and are often confined to specific types of objects. So, how can we clearly interpret data without spending hours of expert time to sequence areas of interest?
![Police operatives observing a CCTV system](/media/tjbkiha2/adobestock_285439784.jpeg?anchor=center&mode=crop&width=378&height=330&format=jpeg&quality=80&rnd=133077098402600000)
The Approach
We’ve developed an innovative system for extracting high-level features from media, negating the need for pre-tagging the data. This approach allows us to search and compare any single image to find similarities, creating a flexible, focused user experience.
![Media Miner logo](/media/cnejxbtd/media-miner-01.jpg?anchor=center&mode=crop&width=378&height=330&format=jpg&quality=80&rnd=133077191685730000)
THE OUTCOME
We developed Media Miner™. This is a tool that captures, analyses and segments huge quantities of video data to find common patterns, scenes and objects against a single image. Developed with a microservices architecture, Media Miner™ is scalable, easy to maintain and independently deployable.
![A padlock on a shield with binary](/media/xbelwsr3/shutterstock_301660784-1.jpg?anchor=center&mode=crop&width=600&height=600&format=jpg&quality=80&rnd=133077270456400000)
Related news, insights and innovations
Find out more about our cutting-edge expertise.
![IMG 20240419 WA0015](/media/fbadfl0a/img-20240419-wa0015.jpg?anchor=center&mode=crop&width=104&height=104&rnd=133579979243670000)
![Ribbon Cutting 3](/media/ymqhc504/ribbon-cutting-3.png?anchor=center&mode=crop&width=104&height=104&rnd=133451331339670000)
![Adobestock 186550922](/media/b5qlc2jt/adobestock_186550922.jpeg?center=0.50802407221664991,0.75689223057644106&mode=crop&width=104&height=104&rnd=133650777722000000)
![An in-platform view of Roke’s AI-powered PRISM sensor system](/media/ansn1vny/people-1.jpeg?anchor=center&mode=crop&width=104&height=104&rnd=133603227007970000)