KBS+ Ventures invests in Computer Vision startup, Dextro
Stanford professor Fei Fei Li said, āvisual data is truly the ādark matterā of the internetā.Ā While mainstream mobile devices enable consumers to easily capture and upload video in a few clicks, other than limited metadata, developers and publishers have no way to analyze, or draw conclusions from this unstructured data.Ā Consumers watch over 6 billion hours of videos on YouTube per month ā 100 hours of content is posted on YouTube alone per minute ā and that volume will continue to increase.Ā Several companies (e.g. Ditto Labs, Image Vision) have developed impressive image recognition technology to provide advanced analytics on photos; however, due to its level of complexity, few technologists have applied computer vision to video analytics. While text-based contextual targeting has been widely accessible to publishers and marketers for years, the rich visual data contained within video has remained a locked box.
We are pleased to announce our investment in Dextro, a computer vision company whose technology enables businesses to search, filter, query, and capture actionable insights from their video and photo content.Ā This platform determines the presence, absence and prominence of any brand logo, object, and scenery in photos or long form video content.Ā Automated visual analysis allows for efficient categorization and online media targeting, addressing the lack of transparency surrounding competitive tracking, product placement, use cases, or campaign engagement. Ā Online video is growing faster than most other advertising formats, and is projected to reach $5B in 2016. Ā SSPs, DSPs, premium publishers, video ad networks and large media companies can leverage Dextro's technology to provide a much more intelligent video ad product through visual-based contextual targeting.












