Collective Behavior: Efficient and Generalizable Object Tracking on Resource-Limited Platforms

Target Audience: Bachelor's and Master's Students

Overview:
Object detection and tracking remain highly active areas of research in the fields of robotics and computer vision. However, contemporary methods often present limitations when applied to resource-constrained platforms, and achieving consistent, adaptable performance across dynamic scene settings poses additional challenges. This initiative aims to address and propose resolution for these issues.

Problem Statement:
The challenge is to enable efficient object tracking on platforms with limited resources, ensuring adaptable and generalizable capabilities across a range of dynamic settings. Achieving this goal requires evaluating machine learning and deep learning techniques to enhance tracking capabilities, finding the optimal balance between requirements in speed and accuracy in various evolving scenarios.

Requirements:
Good knowledge in general computer vision and machine learning, OpenCV, C++ and/or Python.

If you are interested, please contact Carlos Pinheiro.