Human Detection And Tracking

Alternatives To Human Detection And Tracking
Project NameStarsDownloadsRepos Using ThisPackages Using ThisMost Recent CommitTotal ReleasesLatest ReleaseOpen IssuesLicenseLanguage
2 months ago235otherC++
OpenPose: Real-time multi-person keypoint detection library for body, face, hands, and foot estimation
Opencv4nodejs4,72962422 months ago120May 13, 2020290mitC++
Nodejs bindings to OpenCV 3 and OpenCV 4
Node Opencv4,275307615 months ago31March 10, 2020126mitC++
OpenCV Bindings for node.js
Pigo3,992153 months ago24November 02, 20212mitGo
Fast face detection, pupil/eyes localization and facial landmark points detection library in pure Go.
Face Mask Detection1,355
7 months ago20mitJupyter Notebook
Face Mask Detection system based on computer vision and deep learning using OpenCV and Tensorflow/Keras
5 years ago4
A Face detector for anime/manga using OpenCV
Head Pose Estimation1,024
11 hours ago25mitPython
Realtime human head pose estimation with ONNXRuntime and OpenCV.
3 years ago8mitC++
Real time deformable face tracking in C++ with OpenCV 3.
Emotion Detection907
2 months ago12mitPython
Real-time Facial Emotion Detection using deep learning
6 years ago14Python
(WARNING: This repository is NO LONGER maintained ) Real time face detection and recognition base on opencv/tensorflow/mtcnn/facenet
Alternatives To Human Detection And Tracking
Select To Compare

Alternative Project Comparisons

Human detection and Tracking

Say Thanks!


In this project we have worked on the problem of human detection,face detection, face recognition and tracking an individual. Our project is capable of detecting a human and its face in a given video and storing Local Binary Pattern Histogram (LBPH) features of the detected faces. LBPH features are the key points extracted from an image which is used to recognize and categorize images. Once a human is detected in video, we have tracked that person assigning him a label. We have used the stored LBPH features of individuals to recognize them in any other videos. After scanning through various videos our program gives output like- person labeled as subject1 is seen in video taken by camera1, subject1 is seen in video by camera2. In this way we have tracked an individual by recognizing him/her in the video taken by multiple cameras. Our whole work is based on the application of machine learning and image processing with the help of openCV.This code is built on opencv 3.1.1, python 3.4 and C++, other versions of opencv are NOT SUPPORTED.


  • opencv [v3.1.1]

    • Installation in linux: For complete installation of opencv in ubuntu you can refer here.
    • Installation in windows For complete installation of opencv in windows you can refer here
  • python3

    • In Ubuntu python 3.4 can be installed via terminal with the command given below: sudo apt-get install python3
  • python libraries: Here is a list of all the python dependencies

    • Python Image Library (PILLOW)
    • Imutils
    • numpy
  • C++


  • The code follows the steps given below:
    1. First it reads a video and process each frame one by one.
    2. For each frame it tries to detect a human. If a human is detected it draws a rectangle around it.
    3. after completing step 2 it tries to detect human face.
    4. if a human face is detected it tries to recognize it with a pre-trained model file.
    5. If human face is recognized it puts the label on that human face else it moves to step 2 again for next frame
  • The repository is structured as follows:
    • : This is the main python file that detects and recognizes humans.
    • main.cpp : This is the main C++ file that detects and recognizes humans.
    • : This python script is used to create model file using the given data in data/ folder
    • model.yaml : This file contains trained model for given data. This trained model contains LBPH features of each and every face for given data.
    • face_cascades/ : This directory contains sample data for testing our codes. This data is prepared by extracting face images of a praticular person from some videos.
    • scripts/ : This directory contains some useful scripts that we used to work on different problems.
    • video/ : This directory contains some of the videos that we used to while testing.



Don't forget to install the necessary libraries described in the install paragraph above.

First you need to run the file, which uses the images in /data to create a .yaml file

  • In the project folder run
  • To run the python version of the code you have to put all the input videos in one folder and then provide the path of that folder as command line argument:
python3 -v /path/to/input/videos/  

Example- for our directory structure it is:

 python3 -v /video 


  • To compile the C++ version of the code with openCV the command is:
 g++ -ggdb `pkg-config --cflags opencv` -o `basename name_of_file.cpp .cpp` name_of_file.cpp `pkg-config --libs opencv` 

Example- for our directory structure it is:

 g++ -ggdb `pkg-config --cflags opencv` -o `basename main.cpp .cpp` main.cpp `pkg-config --libs opencv` 
  • To run the C++ version of the code you have to put all the input videos in one folder and then provide the path of that video as command line argument:
./name_of_file /path/to/input/video_file 

Example- for our directory structure it is:

 ./main /video/2.mp4
  • creating your own model file; just follow the steps given below to create your own model file:
    • for each individual rename the images as subjectx.y.jpg for example for person 1 images should be named as subject01.0.jpg , subject01.1.jpg and so on.
    • put all the images of all the persons in a single folder for example you can see data\ folder then run this command given below: python3 -i /path/to/persons_images/

Performance of code

  • Since this is a computer vision project it requires a lot of computation power and performance of the code is kind of an issue here.
  • The code was tested on two different machines to analyse performace. The input was 30fps 720p video.
    • On a machine with AMD A4 dual-core processor we got an output of 4fps which is quite bad.
    • on a machine with Intel i5 quad-core processor we got an output of 12fps.


alt text alt text alt text alt text

You can find project report here

To do

  • improve the performance of the code
  • improve the accuracy of the code and reducing the false positive rate.
  • improve the face recognition accuracy to over 90 percent

Special Thanks to:

Popular Opencv Projects
Popular Face Projects
Popular Machine Learning Categories
Related Searches

Get A Weekly Email With Trending Projects For These Categories
No Spam. Unsubscribe easily at any time.
C Plus Plus
Face Recognition
Face Detection