Dual-Camera 3D Tennis Scene Understanding
A multi-model perception system that maps court geometry, players, and fast-moving ball trajectories into real-world 3D coordinates.
Project archive
Search and filter 52 projects spanning computer vision, robotics, machine learning infrastructure, and applied research.
A multi-model perception system that maps court geometry, players, and fast-moving ball trajectories into real-world 3D coordinates.
Developed a grid-guided tennis court keypoint detection model for identifying court line intersection points from camera images.
Developed a YOLO-based person detection model for tennis court videos that predicts both person bounding boxes and additional person-level attributes in a single perception model.
An auto-labeling pipeline that uses SAM-3D-Body outputs to generate person detection annotations for tennis-court scenes.
Developed a temporal tennis ball detection model that uses multi-frame input to detect the ball and predict its diameter directly from the detection head.
A unified temporal model that combines 3-frame tennis ball detection with direct ball radius regression.
A court keypoint localization model for vision-based robot positioning and tennis-court spatial awareness.
A custom 2-frame temporal person detection model for tennis player localization in difficult court-view footage.
A real-time player action recognition system for safety-aware machine control in sports robotics.
A dedicated tennis ball radius regression model for estimating tiny ball scale from cropped detections.
An interactive SAM3-based auto-labeling pipeline for accelerating tennis ball detection dataset generation.
A custom 3-frame temporal object detection model for tennis ball tracking in challenging match footage.
A real-time, edge-powered tennis analytics stack combining player tracking, pose, ball trajectory, and court mapping.
A point-cloud processing pipeline for ground segmentation and obstacle extraction in outdoor robot navigation.
Synthetic tennis-court simulation work in Isaac Sim and Webots for generating ball-tracking datasets and testing camera layouts.
A lawn-care perception component that distills foundation-model masks and classifications into a compact weed detector.
An object detection pipeline for a sprayer robot, from foundation-model auto-labeling to YOLOv11n edge deployment.
A grass quality segmentation system that converts weak labels into patch-level supervision and produces quality heatmaps.
A lightweight charging-station keypoint detector for 3D localization and auto-parking.
A stereo matching pipeline that uses a large teacher model for auto-labeling and a smaller student model for edge deployment.
An end-to-end instance-segmentation workflow spanning foundation-model auto-labeling, training, evaluation, and edge deployment.
A hierarchical quadruped locomotion framework that combines Deep Reinforcement Learning with Model Predictive Control.
Proprioception-only locomotion policies transferred from simulation to a custom quadruped across flat and rough terrain.
Deep reinforcement learning experiments for quadruped locomotion on a Unitree A1 robot.
A keypoint detection system for indoor doors and windows, designed to improve structural perception in indoor robotics environments.
A binary semantic segmentation pipeline for separating a robotic arm from the background.
An RGB-D auto-labeling pipeline that generates 6-DoF grasp-pose annotations for robotic pick-and-place in unseen domestic environments.
A survey and comparison of feature matching methods for visual odometry, visual localization, and SLAM.
A custom human-object interaction dataset and model training workflow.
A survey of 3D multi-person pose estimation methods for recovering human pose in both relative and absolute coordinate settings.
A survey of recent facial expression recognition algorithms and emotion categories.
A survey of human mesh reconstruction methods using parametric body models such as SMPL.
A spatio-temporal action recognition project using contextual relationships between actions, people, and surrounding objects.
A synchronized LiDAR and multi-view pipeline that increased depth-label density from 6% to 58% through calibrated frame fusion.
A self-supervised multi-view stereo reconstruction workflow for generating stronger depth supervision on custom data.
A stereo depth estimation project focused on modeling both aleatoric and epistemic uncertainty for safer real-time depth prediction.
Added real-time aleatoric uncertainty estimation to a semantic segmentation pipeline so uncertain regions can be filtered before downstream use.
An indoor semantic segmentation project comparing transformer-based and CNN-based segmentation models on custom robotics data.
An indoor 3D perception pipeline combining stereo depth, 3D object detection, and 6D pose estimation.
A human attributes recognition system focused on safety-related visual attributes and edge deployment.
A real-time person re-identification system for intelligent video surveillance across six adjacent-street cameras.
Master's thesis on deep discriminative feature learning for person re-identification.
Explored Generative Adversarial Networks for synthetic image generation using the classic generator-discriminator training setup.
Contributed to an EEG-based system for mental state and emotion classification.
Contributed to the graphics software for a smart multi-function display using OpenGL.
A computer vision system for measuring shooting target accuracy and managing shooter records.
A single-cycle processor designed in Verilog for a computer organization and architecture course.
An embedded system that extracts navigation information from a GPS receiver using an 8051 microcontroller.
A hardware-based digital dice circuit that generates random dice values and displays them on a 7-segment display.
An 8051 microcontroller-based distance measurement system designed as a movement orientation aid for dark or low-visibility environments.
A Java Sudoku game developed for an object-oriented programming course, with a graphical interface and event-driven interaction.
A compact C-language Tic Tac Toe game built as an Introduction to Computer Programming course project.
Try a broader search or a different category.