COMPUTER VISION: OBJECT TRACKING USING SELF-BALANCING TREE FOR OPTIMIZED DATA PROCESSING

Tracking Multiple Objects and Prioritizing them can be an expensive operation. In a real-world scenario where, for example, many cars are being tracked coming through a highway by a certain characteristic such as car color this leads to one needing to process them efficiently, especially when the cars are quickly coming in and out of the screen. This research aims to see if there is an optimal way to process and sort many tracked objects efficiently; thus, the purpose of this Thesis Research Project is to answer the central question of: “whether the AVL self-balancing Binary Search Tree can be used for optimization of data processing from a Computer Vision Robotics camera source?”

The Tree data structure is used in many areas of Computer Science and Software Engineering; there are different variations, each with different characteristics for various applications. This research uses the AVL (Adelson-Velsky and Landis) self-balancing binary search tree to see differences in Asymptotic Runtime for insert and traversal operations. A number of specialized hardwares and softwares will be used to conduct this research with rigorous test benchmarking to get the most accurate results. The results of this Thesis Research Project shows the AVL tree being stable with the complexity running time maintaining O (logn) for the tested operations with varying batch data sizes making it feasible for post-processing use in Computer Vision and other industry applications.