Object detection and tracking

Position - Software Developer at AitoeLabs
Technical skills - C++, OpenCV, Python, Tensorflow, PyTorch

At AitoeLabs, I worked with a team of engineers to build efficient video analytic solutions for internal security. To ensure ATM security we developed deep learning models and algorithmic codes on the live video stream.
  I was affiliated with the object detection and tracking group. Our implemented human detection algorithm’s performance is displayed in below.

atm surveillance
Figure: Detecting multiple objects (people) in a live stream

  Object tracking deals with tracking detected objects by assigning IDs to them. A theft or vandalizing of an ATM generally happens when there is a group of people (more than the allowed number) in the ATM or if a person is loitering around inside the ATM for a prolonged duration. Using object tracking methodology we developed algorithms to detect this kind of behavior and on detection alerted the security personnel/ company.

Dimple Bhuta
Dimple Bhuta
Principal Engineer – State Estimation and Control

My research interests include robotics, computer vision and bio-inspired design.