Published Research Paper
Published Research Paper in the Curieux Academic Journal investigating the accuracy of various object detection algorithms (e.g. YOLOv5) in detecting a hockey ball. Trained AI models on a dataset of 350 images to detect hockey balls with over 85% accuracy in match situations.
Authored Extended Essay
Analyzed the performance of object detection models. I compared non-neural approaches, such as SIFT, with neural (machine learning) based approaches to ascertain their ability to detect common household objects. My results showed that neural approaches to computer vision improved precision and decreased false positive rates.
