For face recognition, medialoopster uses the on-premise AI engine ReCAP and the Cloud AI services provided by aiconix. Customers can decide whether they want to run the services on-premise or in the cloud.

The company explains that data provided by the engines for face and person recognition is displayed frame accurately in medialoopster so that people can now be searched specifically. In the video player, the recognized persons are marked with bounding boxes and names. The colored sections below the player timeline provide visual feedback on the time periods in which the respective person is included in the video. In addition, persons found in the respective video are represented in round pictures in this area.

Existing capabilities for automatic transcript generation are expanded and a generic option for exporting subtitles using the software’s XML processor is added. In addition, transcript metadata has been more tightly integrated with medialoopster metadata management, ensuring that transcripts can be inherited when exporting files from Adobe Premiere Pro.

The ability to display different aspect ratios for video assets in different social media formats is improved, as is multi-site workflow for asset sharing in different medialoopster locations. In future it will be possible to share an entire collection, such as the content of favorite lists.