Long-form video understanding
We study how characters, locations, actions and ideas develop through a complete video. The difficult part is retaining context over minutes or hours.
Visonic Research
Research
Visonic is a research-led AI company. We work on long-form video understanding, computer vision and machine cognition.
Research areas
Understanding video over time
Computer vision
We identify people, objects, actions and on-screen text, then connect those visual signals through space and time.
Machine cognition
We study how a system can connect perception, memory and reasoning to determine what happened and which details matter.
Applied research