EVEEX - Experimental Video Encoder ENSTA Bretagne

Introduction

This project is maintained and currently developed by a team of 5 students at ENSTA Bretagne.
This project is a proof of concept (POC) for the whole project that consist in making a embedded video encoder and decoder on two different FPGA. The end goal is to have a first FPGA with a camera attached that compresses the video flux and send it to another FPGA that can decode the data and show the content on a screen.

Features

• Generate test images from a json description file (WIP)
• Encode an image to a Bitstream
• Decode a Bitstream into an image (WIP)
• Send Bitstreams from a client to another

Installation

Before you begin, make sure you have a minimal python 3.8 installation on your system.

The first step is to clone the project and go into the project's folder.

git clone https://github.com/EVEEX-Project/EVEEX-Code.git
cd EVEEX-Code/python_prototype

The next step is to make sure you have all requirements already installed. You will need the pip tool that is often already installed with python (if not please check the documentation for the extra steps).

python -m pip install -r requirements.txt

This line will install every requirements in the text file in order to be able to launch the project.

Usage

By simply typing python eveex.py you will be greeted with an helper message letting you know which command is available.

Image generation from JSON

There is an option to generate a blank image or a mosaic image with the tools provided. We went one step further and created a way to generate our own test images from a JSON file.

To do so you need to create a description JSON file with the following starting content.

{
  "header" : {
    "size" : [100, 100],
    "background_color": "black"
  },
  "content" : []
}

This code will create an image of size 100 by 100 pixels with a black background color. You may notice the empty array named "content". This array will contain every object that should appear on the image.

Please note that in order to be drawn on the canvas, a shape cannot overflow the canevas

For example if we want to put a red rectangle on the canevas, we have to add the following code into the content array.

{
    "type" : "rectangle",
    "position" : [10, 20],
    "size" : [20, 20],
    "color" : "red"
}

The required arguments are the shape's type, the position of the top-left corner, the size of the rectangle and it's colour.

If we want to draw a green circle we need to add the following to the array.

{
    "type" : "circle",
    "position" : [80, 10],
    "size" : 4,
    "color" : "green"
}

With a bit of tinkering one can generate complex shapes and great test images for the encoder like the following example.

Encoding an image

One can generate an image from a JSON description file for example. Let's save this image as test.png.

In order to encode the image and save it as a bitstream file called bitstream.txt one can use the following command.

python eveex.py -e test.png bitstream.txt