1) Developed a YOLOv8 object detection model with impressive accuracy and improved segmentation using YOLOv8-seg and SAM.2) Integrated Co-Tracker and created custom algorithms for precise bat tracking.3) Implemented YOLOv8-pose, achieving significantly faster processing while maintaining accuracy.4) Evaluated methods across DRS angles, achieving high accuracy on a diverse test set of video frames.

Engaged in the development of combinatorial optimization algorithms tailored for UAV path planning amidst obstacles, with a specific focus on the Bees Algorithm. Conducted an in-depth analysis of various performance heuristics employed in optimizing using the Bees Algorithm, drawing comparisons with analogous nature-inspired optimization algorithms. Executed simulations on MATLAB to showcase the analysis and resulting insights. Presented this work at the 2nd International Workshop on the Bees Algorithm and its Applications in 2022.

Implemented Bees Algorithm (2-parameter version) for efficient hyperparameter exploration in deep learning models, avoiding brute force techniques.Developed a specialized CNN+LSTM-based deep learning model for the IEEE 2012 Prognostic Challenge, focusing on predicting the remaining useful life of ball bearings (5 training bearings, 11 test bearings).Co-authored a book chapter titled "Bees Algorithm for Hyperparameter Search in Deep Learning Models," exploring the algorithm's application in enhancing deep learning hyperparameter search processes.Presented research findings at the "2nd International Workshop on the Bees Algorithm and its Applications (BAA-2022)," facilitating discussions and knowledge exchange within the research community.