Services > Artificial intelligence > Image recognition
AI-driven image recognition and analysis
AI-driven image recognition processes visual content quickly, accurately and at a scale that manual review simply cannot match. We can integrate image analysis into your existing platforms to automate processes such as categorisation, quality issue detection and data extraction from large numbers of images.
The applications are broad. From automating quality assessments to making image and video content searchable and discoverable, image recognition reduces the need for expensive and time consuming manual review, speeds up workflows and provides insights that would be impractical to gather any other way.
How we’ve used AI-driven image recognition
Automated photo analysis
We integrated AI image analysis into the Football Foundation's PitchPower app to automate the assessment of grass, weed and earth coverage from user submitted photographs. The tool improved categorisation accuracy from 50% to 85% and dramatically reduced the time spent manually reviewing tens of thousands of images, speeding up the funding approval process for football clubs across England.
Video content analysis and categorisation
For Microsoft we used AI to analyse and categorise lengthy product demo videos, breaking them down into searchable segments based on their visual and spoken content. The categorised content fed an AI personalisation system that read user intent and behaviour, surfacing the segments most likely to interest each viewer. Users could jump straight to the relevant section rather than watching the full video, while Microsoft gained insight into which content was most in demand, helping to inform their creative process.