Sixth Sense Retrieval will demonstrate a complete workflow for image recognition in news videos, unlocking a wealth of archived content, streamlining content production and delivering higher-quality videos to audiences.
User demand for moving images is increasing, but this material is not easily searchable. Existing content management systems rely on short text descriptions for later retrieval. This manual annotation is labour-intensive and cannot cover all images in detail or predict future searches. The prototype tool will use the latest advances in the fields of image segmentation, object detection and scene summarisation to automatically analyse video content, helping journalists search for relevant footage using plain language.
Google serves cookies to analyse traffic to this site. Information about your use of our site is shared with Google for that purpose. Find out more.