How SVEN Works
The SVEN AI-Video-matic:



Step 1

Video Camera Input

Live video from the surveillance camera is read both into the computer vision software as well as the effects processing software MAX/ MSP Jitter.



 Camera hangs high above SVAN              The lonely streets                                      SVEN awaits potential stars                   The city awakens, unsuspecting

Step 2


Computer Vision
software: C++/ Open CV

Vincent's computer vision software uses live video data and tracks any person present in the video. It can distinguish facial expressions, clothing color, and body segmentation - among other things. This information is sent to MAX/MSP Jitter.

                     Tracking a pedestrian and reporting her hair and shirt color                         Information gathered before effects (next step) are added

Step 3


Video Processing
software: MAX/ MSP Jitter

Wojciech's MaxMSP/ Jitter patch matches the computer vision data to the database of music videos. It then processes the video footage, performing music video cinematography, editing, and special video effects.




Facial Expression Matches:

Jack White's nemesis Doin' the Bono squint Ashcroft's surprised twin

Effects Processing:

                                    SVEN interprets U2's "Vertigo" (original shown below)

Step 4

Video Output

The final product is displayed on an external monitor







Home | How it Works | Live Shows | About SVEN | SVEN CV Software