Using a Stereo-Vision Setup to reliably track a Player's Elbow

A simple project using MediaPipe and OpenCV to detect human body landmarks from video. Includes both 2D pose overlay on video and real-time 3D skeleton visualization with Matplotlib.

3D pose estimation of mushrooms from pointcloud - trained on synthetic data

Multiperson Human Pose Estimation on a video in 3D

🌊 Simplify configuration with VIBE, a readable, fast-format that eliminates complexity while enhancing clarity and structure in your development workflow.

This repository provides a solution for estimating 3D positions using feature-based detection and deep learning model integrated with OpenCV's Deep Neural Network (DNN) and the Robot Operating System (ROS).

Undergraduate Internship 학부인턴십 활동 (July 2022~)

Pytorch implementation for the paper: "DAGM-Mono: Deformable Attention-Guided Modeling for Monocular 3D Reconstruction"

Exploring the use of Adversarial Constrained Autoencoder Interpolation (ACAI) to improve the quality of latent space for 3D human pose representation using the h36m dataset.

Pytorch implementation of 3D human keypoints estimation models

The official repository of the paper "X as Supervision: Contending with Depth Ambiguity in Unsupervised Monocular 3D Pose Estimation"

3D Markerless Tracking Toolbox

Official implementation of CVPR2020 paper "VIBE: Video Inference for Human Body Pose and Shape Estimation"

New egocentric synthetic dataset for egocentric 3D human pose estimation

This is the official code for calibration in multi-hypothesis human pose estimation

EventEgo3D++: 3D Human Motion Capture from a Head-Mounted Event Camera [IJCV]

Official project website for the AAAI 2022 paper "Stereo Neural Vernier Caliper"

Reference ImageNet implementation of SelecSLS CNN architecture proposed in "XNect: Real-time Multi-person 3D Human Pose Estimation with a Single RGB Camera". Also includes code for pruning the model based on implicit sparsity emerging from adaptive gradient descent methods.

A simple yet effective method for 3D pose estimation using only 2D supervison

This is a for of the original implementation, sharing so that anyone interested in using it on windows without a gpu in the system can run it