Inspiration
Can we keep check on the carbon footprint and take steps to reduce pollution? Can we solve the parking problem and suggest areas to park in real time? How To Create and maintain, benchmark datasets? Can we provide real time video analytics to for traffic congestion and suggesting other paths? Can we eliminate human intervention in monitoring ?
What it does
VisionCV : Deep learning based Real time Intelligent Surveillance High performance Yolov Architecture and Maintenance(VisionCV).
How we built it
Input: Live Video Stream Models: YOLO-V3 based Deep learning architecture Object detection {Humans, vehicles, luggages, trolleys} DeepSORT Algorithm Activity recognitions: {Running, Walking, Carrying} Output: Automatic Alert count of objects in a given area.
Challenges we ran into
Accomplishments that we're proud of
SAAS platform reduces cost by 60% Human intervention is minimized and the solution is automated A modular, scalable server Architecture for ‘storing the imagery and video data Alerts if there is carbon footprinting above given threshold & do path redirections accordingly. Paths with less traffic are suggested, thus saving lot of time for commuters
What we learned
What's next for VisionCV
Financial Support for establishing GPU based High performance Hardware setup for running Deep Learning models. High Speed internet Connectivity. Funds to work on the proposed approach. Huge amount of imagery data generation for OFC network elements and pre-processing or Bench marked dataset. Video surveillance data for validation Platform for validation.
Log in or sign up for Devpost to join the conversation.