Automatically run an analysis of your video content which returns image assets of extracted faces together with name labels

Using a video assets residing within your VidiCore environment, the Face and Label Extractor is able to analyze and extract faces detected in the content and group them together with name tags identified from lower third. The extracted keyframes containing the identified faces will then be imported as new image assets in VidiCore. With extracted faces together with metadata describing the person in the picture you are then able to train face recognition models, and analyze your video content library to detect which time spans in the content the person is appearing.

For all customers that are looking beyond the pre-trained face detection models offered and instead wants to build their own model people such as regional politicans or local celebrities. This is a valuable tool for generating source material to be used together with face recognition services that can be trained, such as the Face Training and Analysis Powered by DeepVA

Seamless integration into any VidiCore system, anywhere.

Requires VidiCore version 21.3 or later.