Detecting moving objects by computing differences between consecutive frames, dilation for connecting nearby contours, and non maximal suppression to filter overlapping bounding boxes.

This code uses feature detection and matching to detect a target object. Old concept, but still pretty cool to see run in almost real-time.

I know, boring.. but why not? Using the sliding window approach and perspective transformation to detect and highlight lane lines from dashcam.

This project aims to track the camera's motion and estimate its pose using ORB feature detection, and reconstruct the 3D environment.

An open source video surveillance application to offer an alternative to high-cost video surveillance systems. Provides real-time video streaming, email notifications, AI human detection, basic authentication, and a containerized environment for easy deployment to any platform with a USB camera.

This project aims to predict successful shots from basketball videos. It uses YOLOv8 to detect the ball/rim and kalman filtering + polynomial regression to track the ball and predict its trajectory towards the rim

Exploring camera pose estimation applications using a camera from a mobile phone

This project aims to implement Reinforcement Learning (RL) to enable the Boston Dynamics' Spot robot to autonomously navigate in low-gravity environments.

This was my UG senior design project: a landmine-detecting robot. Complete integration, from hardware to software. It uses ML for landmine detection and lidar for navigation. It also features a GUI showing robot location, detected mines, waypoints, and control buttons.

My personal responsive website built with plain javascript/css/html, bootstrap, and jquery.

Fully 3D printed 6DOF Robot Arm, developed using Arduino.