Sign2Speech RICM4 2015-2016 User Manual

How to run the project

The project is splitted into two parts:

• the gesture recognition software
• the chat application

Run the gesture recognition software

In order to run this part you only need a Windows computer with the RealSense Camera connected and the chat server running.

A dictionary in JSON file format has to be present in the same directory as the runnable file (Sign2Speech.exe). Check the github repository and the output.json file to understand the architecture of the dictionary.

The main thing to know about the software is that you can run it with some parameters detailled with the "-h" parameter. So, if you're running the server on another machine you will have to set the address parameter with the "-a" parameter.

The websocket and the gesture recognition will only work if the software can connect to a room in the chat server. By default this room is named "test".

Once the software is running, you can sign and test the application, and check messages in the chat.

If you try to use the project in Visual Studio, you will probably have to add the props file for RealSense SDK. The props file is under the props folder and should be added to all the project in the Properties Manager.

Run the chat server

For this part you will need to have Maven 2 installed on your computer. It can be done both on Windows and Linux.

The server uses a framework called Wisdom Framework. This technology embeds some very usefull features that can help you in your development.

Their are two way to run the server:

• if you are programming on the server, run it with the command: mvn clean wisdom:run -dskipTests in the directory where pom.xml is(-dskipTests is optionnal and only reduce the time of execution by passing tests, if you're running tests in your app remember to remove this parameter)
• if you want to deploy on a distant machine, run: mvn clean install on your computer in the directory where pom.xml is. After that copy the file sign2speech-webrtc-1.0-SNAPSHOT.zip on the distant machine and extract. In order to finish the running process, run with: ./chameleon.sh or ./chameleon.bat if you are on Windows. The server will start on the port 9000, but by changing http.port in the conf/application.conf you can run it on the port you want (80 for example).

For more informations about Wisdom Framework and how it works, please refer to its documentation: Wisdom Framework