JPEG Compression Stage Viewer

  

Welcome to the repository dedicated to our Multimedia project during our undergraduate studies, focusing on the exploration of the JPEG compression method. This interactive program provides a detailed view of the codec, allowing users to customize their experience through an intuitive scripting language.

Key Features:

Interactivity:

• Explore each phase of the JPEG codec interactively.
• Instruct the program with a simple scripting language.

Detailed Visualization:

• Visually observe the impact of each stage on image quality.
• Combine stages to understand compression decisions.

Side-by-Side Comparison:

• Compare various stages of the JPEG codec for a comparative analysis.
• Gain insights into different compression settings.

---

---

Contents

mult_tp1
├── ARCH
├── DEV
│   └── test
│   └── img
├── PM
│   ├── docs
│   └── group
├── PROC
├── PROD
├── QA
│   └── docs
└── REQ

Requirements

Python 3.8 or above

Libraries:
python -m pip install -r requirements.txt

Usage

usage: main.py [-h] (-i PATH | -a PATH) [-c CHANNEL] [-m     ] [-n NAME] [-e] [-y | -r]
               [-p PADDING] [-s  ] [-d DCT] [-q QUANTIZE] [-f]

optional arguments:
  -h, --help            show this help message and exit
  -i PATH, --image PATH
                        show the image on the given path
  -a PATH, --config PATH
                        use a configuration file with commands to run multiple plot instances
  -n NAME, --name NAME  give a name to the plot
  -e, --encode          encode image using JPEG codec and display steps
  -y, --ycbcr           convert the image channels to the YCbCr color model
  -r, --rgb             convert the image channels to the RGB color model (default)

  -c CHANNEL, --channel CHANNEL
                        select a color channel: {1, 2, 3}
  -m      , --colormap
                        provide a colormap for the image

  -p PADDING, --padding PADDING
                        add padding to the image unitl a certain value
  -s   , --downsample
                        downsample the image by a given set of values {0, 1, 2, 4}
  -d DCT, --dct DCT     calculate the dct of the image (must be multiples of 8, 0 to apply on
                        the whole channel)
  -q QUANTIZE, --quantize QUANTIZE
                        quantize the image with a defined quality factor [0,100]
  -f, --dcpm            encode the DC coeficients of the image

Scripting the behaviour of the program

A configuration file in a basic scripting language can be used to visualize different steps of the compression algorithm side by side.

To use: main.py -a path/to/config_file

Example of a configuration file structure:

plot "RGB Channels"
-i "img/barn_mountains.bmp" -m 0 0 0 1 0 0 -c 1 -r -n "R Channel"
-i "img/barn_mountains.bmp" -m 0 0 0 0 1 0 -c 2 -r -n "G Channel"
-i "img/barn_mountains.bmp" -m 0 0 0 0 0 1 -c 3 -r -n "B Channel"
end
plot "YCbCr Channels"
-i "img/barn_mountains.bmp" -m 0 0 0 1 1 1 -c 1 -y -p 32 -n "Y Channel"
-i "img/barn_mountains.bmp" -m 0 0 0 1 1 1 -c 2 -y -p 32 -n "Cb Channel"
-i "img/barn_mountains.bmp" -m 0 0 0 1 1 1 -c 3 -y -p 32 -n "Cr Channel"
end
plot "YCbCr Channels whith 422 Downsampling"
-i "img/barn_mountains.bmp" -m 0 0 0 1 1 1 -c 1 -y -s 4 2 2 -p 32 -n "Y Channel"
-i "img/barn_mountains.bmp" -m 0 0 0 1 1 1 -c 2 -y -s 4 2 2 -p 32 -n "Cb Channel"
-i "img/barn_mountains.bmp" -m 0 0 0 1 1 1 -c 3 -y -s 4 2 2 -p 32 -n "Cr Channel"
end
plot Original
-i "img/barn_mountains.bmp"
end

Flags

Command	Arguments	Action
-i	PATH	Selects an image file to use
-m	FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT [0,1]	Creates a color map for image visualization
-c	INT {1,2,3}	Selects one of the image channels
-y		Selects YcbCr color mode
-r		Selects RGB color mode
-s	INT INT INT {0,1,2,4}	Subsamples image channels according to a configuration
-n	STRING	Name of the image subplot
-p	INT	Adds padding to the provided image
-d	INT	Performs DCT on blocks of the given size or on the entire image if not provided
-q	INT	Quantizes the image using quantization matrices
-f		Applies DPCM encoding
-e		Encodes the image and displays intermediate steps
plot	STRING	Initiates a plot code block with the provided name in STRING
end		Ends a code block