Our client is developing a mental health screening tool targeted for adolescents. This tool will allow physicians to track the mental health state of patients through the use of video. Our client needed a platform to explore and process incoming video data in order to produce features and insights that physicians can use to track mental health. Through the tool’s proposed use of gamification and survey questions, each video is potentially capturing a wealth of information about a patient’s mental health that needed to be automatically extracted in a scalable and secure manner.
We customized and productionized our VIVO platform to enable our client to process videos through VIVO to produce readily available features for analysis, such as emotion, voice, and speech analytics. Our client is able to upload videos to the cloud, which automatically triggers the VIVO platform. Users can rapidly explore VIVO results through a simple application, and download and visualize the features produced for a given frame of the video. We incorporated a third party speech-to-text service in order to optimize the accuracy and security of the transcription pipeline. Finally, the VIVO platform was backed by a Kubernetes cluster, which allows our client to automatically scale up the batch processing of videos whenever necessary.