Computer Vision SS 2020 Challenge

Distinction of foreground and background of a given scenery, the ability to eliminate unwanted elements of this sequence and optionally creating a background video.

• For a full description of the project, please read the project documentation included in the repository:

  https://github.com/Yaccoub/Computer_Vision_Challenge/tree/master/doc

• To submit bug reports and feature suggestions, or track changes:

  https://github.com/Yaccoub/Computer_Vision_Challenge/issues

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Clone this repo to your local machine using:

git clone https://github.com/Yaccoub/Computer_Vision_Challenge.git

• The ChokePoint dataset used for this project can be found under

  http://arma.sourceforge.net/chokepoint/#download

• The folder naming in this dataset follows the pattern {P#*_S#_C#} :

  <Portal.no.type_Sequence.no_Camera.no>

• Each folder contains:

  all_file.txt --> list of all images

  bg_img.txt --> list of background images

Prerequisites

This project requires the following modules:

• Matlab Release 2020a (https://de.mathworks.com/products/matlab.html)
• Image Processing Toolbox (https://www.mathworks.com/products/image.html)

Configuration

• The user can configure the source paths and the program parameters (Left, Right, Start and N) using:

  • Graphical User Interface: start_gui

    Created using the Matlab built-in app developemnt framework App Designer, this user friendly GUI provides standard components such as buttons, check boxes, trees, and drop-down lists to configure and customize the program intuitivly.

  • config.m

    Users can manually configure and customize the program parameters by changing the variables values in the config.m file.

Deployment

After installing the prerequisites and setting up the enviroment you can deploy the program using the command window:

start_gui

This command will open the GUI where you can generate a masked output stream and an extra video with a virtual background

Additional Features

Alternatively to a virtual background picture, the user have the the possibility to use a virtual background video. To use this feature please enter a video as background and check the box background video

Versioning

We use Git for versioning. For the versions available, see the tags on this repository.

Authors

• Alaeddine Yacoub - [email protected] -
• Kheireddine Achour - [email protected] -
• Mohamed Mezghanni - [email protected] -
• Oumaima Zneidi - [email protected] -
• Salem Sfaxi - [email protected] -

License

This project is licensed under the Chair For Data Processing Technical University of Munich