GV

GeezerVison

The World Wide Web, restored to 800 pixels of glory.

Home|About|Gaming cabinet|Archives|How it works|Starter sites

On this site

  • Sign in
  • ActionsCI/CD and automation
  • CodespacesCloud development environments
  • All Features
  • Documentation
  • Training
  • Certifications
  • Build with Copilot
  • Customize Copilot
  • Integrations
  • Trending repos
  • GitHub Sponsors
  • Developer articles
  • Enterprises

This view

  • Original page
  • New address

GeezerVison is showing github.com — GitHub - yasenh/sort-cpp: C++ version of SORT: Simple online and realtime tracking of multiple objects in a video sequence

BranchesTags
Open more actions menu

Latest commit

 

History

77 Commits
77 Commits

Folders and files

NameName
Last commit message
Last commit date
data
data
 
 
docs
docs
 
 
include
include
 
 
src
src
 
 
.gitignore
.gitignore
 
 
CMakeLists.txt
CMakeLists.txt
 
 
Dockerfile
Dockerfile
 
 
LICENSE
LICENSE
 
 
README.md
README.md
 
 
docker_run.sh
docker_run.sh
 
 

Repository files navigation

Introduction

C++ implementation of SORT: Simple, online, and real-time tracking of multiple objects in a video sequence.

Kuhn-Munkres (Hungarian) Algorithm in C++ is forked from: https://github.com/saebyn/munkres-cpp

Dependencies

  • Ubuntu 16.04
  • Docker 18.09.4
  • OpenCV 3.4.2
  • Boost 1.58.0

All of the 3rd party libraries are included in the provided docker image

Build Docker Image

  1. Open Dockerfile, change line #19 ARG USERNAME to your host user name (echo $USER)
  2. Open a terminal and run:
    $ cd /path/to/sort-cpp
    $ docker build -t sort .
    $ ./docker_run.sh

Demo:

Screenshot-1

To run the tracker with the provided detections and visualize the results:

  1. Download the 2D MOT 2015 benchmark dataset
  2. Create a symbolic link to the dataset
    $ ln -s /path/to/MOT2015_challenge/data/2DMOT2015 /path/to/sort-cpp/mot_benchmark
  3. Run the demo
    $ cd /path/to/sort-cpp
    $ mkdir build && cd "$_"
    $ cmake .. && make
    $ cd /path/to/sort-cpp/bin
    # Without display
    $ ./sort-cpp
    # With display
    $ ./sort-cpp -d

Evaluate Metrics

Using the Python implementation of metrics for benchmarking multiple object trackers (MOT) to evaluate metrics.

Dataset Structure

Layout for ground truth data
    <GT_ROOT>/<SEQUENCE_1>/gt/gt.txt
    <GT_ROOT>/<SEQUENCE_2>/gt/gt.txt
    ...

Layout for test data
    <TEST_ROOT>/<SEQUENCE_1>.txt
    <TEST_ROOT>/<SEQUENCE_2>.txt
    ...

Example:
mot_benchmark
├── test
│   ├── ADL-Rundle-6.txt
│   └── ADL-Rundle-8.txt
└── train
    ├── ADL-Rundle-6
    │   └── gt
    │       └── gt.txt
    └── ADL-Rundle-8
        └── gt
            └── gt.txt


Sequences of ground truth and test will be matched according to the `<SEQUENCE_X>`
string.

Example

# Optional for virtualenv
$ source ~/env/bin/activate
$ pip install motmetrics
# Usage
$ python -m motmetrics.apps.eval_motchallenge --help
# Format: python -m motmetrics.apps.eval_motchallenge groundtruths tests
$ python -m motmetrics.apps.eval_motchallenge mot_benchmark/train output/

Result

Screenshot-1

FPS is around 1900 with Debug build.

References

  1. https://github.com/abewley/sort
  2. https://github.com/mcximing/sort-cpp
  3. https://github.com/saebyn/munkres-cpp

About

C++ version of SORT: Simple online and realtime tracking of multiple objects in a video sequence

Resources

Readme
GPL-3.0 license
Activity

Stars

178 stars

Watchers

3 watching

Forks

32 forks
Report repository

Releases

Packages

Contributors

Languages

You are visitor 01402503 — GeezerVison does not store the pages you look up.
Best viewed with Netscape Navigator 4.7 or Internet Explorer 6 at 800×600. No cookies required.