Step 1
Video Intake​
Intake video of conversation between patient and provider
Step 1
Step 2
Feature Extraction​
Separate the video into three modalities - frames, audio, transcript and extract features ​
Step 2
Step 3
Feature Analysis​
Analyze features for each modality using neural networks and proprietary algorithms​
Step 3
Step 4
Modality Fusion​
Use advanced fusion techniques to predict level of mental illness (score, severity)
Step 4
Step 5
Insights​
Provide insights about overall mental state​
Step 5

So how does it work?

Aiberry’s end-to-end architecture and processing pipeline is hosted in the AWS cloud. The entire process from video intake to insight is automated and the results are available immediately.