wegiangb

ml-neuman

Official repository of NeuMan: Neural Human Radiance Field from a Single Video (ECCV 2022)

OCR-SAM

Combining MMOCR with Segment Anything & Stable Diffusion. Automatically detect, recognize and segment text instances, with serval downstream tasks, e.g., Text Removal and Text Inpainting

Deep-Learning-MNIST---Handwritten-Digit-Recognition

This project demonstrates Handwritten digit recognition using Deep Learning

Scoreboard-webcam-OCR

Scoreboard OCR with a webcam and telephoto lens to read digits in real time from a in-venue scoreboard.

aq-biascorrection

Bias correction of air quality CAMS model predictions by using OpenAQ observations.

FootballPassPrediction

Football/Soccer Pass Receiver Prediction using Object Detection/Graph Neural Networks (GNNs)

draw-on-stream-telestrator

Telestrator tool to easy draw on your stream without having to capture your full screen

AvatarCLIP

[SIGGRAPH 2022 Journal Track] AvatarCLIP: Zero-Shot Text-Driven Generation and Animation of 3D Avatars

yolov7-object-tracking

YOLOv7 Object Tracking Using PyTorch, OpenCV and Sort Tracking

predsim_tutorial

TransPose

A real-time motion capture system that estimates poses and global translations using only 6 inertial measurement units

PIP

A real-time system that captures physically correct human motion, joint torques, and ground reaction forces with only 6 inertial measurement units

Track-Anything

Track-Anything is a flexible and interactive tool for video object tracking and segmentation, based on Segment Anything, XMem, and E2FGVI.

Fitness-Tracking-App

Football-Ball-Detection-using-YOLOv5-model

YOLOv5 🚀 is a family of object detection architectures and models pretrained on the COCO dataset. with the weights adjusted to detect the ball in a soccer game.

Dataset-for-Soccer-Action-Recognition-by-PARHN

Four action types, Shooting, Giving pass, Receiving pass and Goalkeeper Diving, in soccer are recognized by introducing pose-projected action recognition hourglass network (PARHN).

FabBits

get interesting bits from videos!

INT_HMR_Model

Capturing the Motion of Every Joint: 3D Human Pose and Shape Estimation with Independent Tokens. ICLR2023 (spotlight)

joint_angles_calculate

Calculate the joint angles of a body pose

MocapNET

We present MocapNET, a real-time method that estimates the 3D human pose directly in the popular Bio Vision Hierarchy (BVH) format, given estimations of the 2D body joints originating from monocular color images. Our contributions include: (a) A novel and compact 2D pose NSRM representation. (b) A human body orientation classifier and an ensemble of orientation-tuned neural networks that regress the 3D human pose by also allowing for the decomposition of the body to an upper and lower kinematic hierarchy. This permits the recovery of the human pose even in the case of significant occlusions. (c) An efficient Inverse Kinematics solver that refines the neural-network-based solution providing 3D human pose estimations that are consistent with the limb sizes of a target person (if known). All the above yield a 33% accuracy improvement on the Human 3.6 Million (H3.6M) dataset compared to the baseline method (MocapNET) while maintaining real-time performance

mediapipe_pose_compare

Joint angle comparison of mediapipe prediction results bvh conversion with ground truth bvh

video2bvh2.0

https://github.com/Dene33/video_to_bvh but with python 3 and tensorflow2.0

VideoTo3dPoseAndBvh

Convert video to the bvh motion file

analytics-handbook

Getting started with soccer analytics

design-guide-for-geospatial-data-portals

awesome-air-quality

An awesome list of air quality resources.

Crowd-Emotion

Emotional sounds of crowd: spectrogram-based analysis using deep learning

moviepy

Video editing with Python

facefun

ResNet-LSTM-GCN

Code for Deep-learning Architecture for Short-term Passenger Flow Forecasting in Urban Rail Transit