Distributed Processing

A distributed image processing system, operating via a console application, in C++. The premise of this application is to showcase the usage of multithreading, networking, and modern C++ standards. The application operates via the conjunction of two executables, server and client. The server is intended to be able to host multiple clients simultaneously.

The original repository for this project can be found here.

Setup

1. Clone the repository.

git clone git@github.com:Khthonian/DistributedProcessing.git

2. Make the build directory.

# Change into repository directory
cd DistributedProcessing/

# Make build directory
mkdir build

3. Generate build files.

# Change into build directory
cd build/

# Generate build files
cmake ..

4. Compile the executables.

make

The executables will be generated to a separate directory, bin.

Necessary Packages

The following packages were necessary to run the application during production:

• fmt
• glew
• opencv
• vtk

You can install these packages with the following command (specific to Arch Linux):

sudo pacman -S fmt glew opencv vtk

Usage

Valid Parameters

A total of 14 filters are included in this implementation:

• Resize filter
• Rotate filter
• Flip filter
• Colour adjustment filters:
  • Brightness filter
  • Contrast filter
  • Gamma filter
• Colour conversion filters¹:
  • BGR to RGB filter
  • BGR to HSV filter
  • BGR to greyscale filter
  • BGR to YCC filter
  • BGR to HSL filter
• Smoothing filters:
  • Gaussian blur filter
  • Box blur filter
  • Sharpening filter

The corresponding inputs are:

Filters	Operation Input	Parameter Input
Resize	resize	Value > 0 to scale.
Rotate	rotate	Value ideally between -360 and 360.
Flip	flip	0 to flip on x-axis, 1 to flip on y-axis, -1 to flip on both.
Brightness	brightness	Any value between 0 and 1 to decrease, any value over 1 to increase.
Contrast	contrast	Any positive value to increase contrast, any negative value to decrease contrast.
Gamma	gamma	Any value greater than 1 to increase, any value between 0 and 1 to decrease.
Colour conversions	colour	rgb, hsv, grey, ycc, or hsl respectively.
Smoothing filters	smooth	gauss, box, or sharp respectively.

Example Usage

Ideally, the application should be used in conjunction with a terminal multiplexer, e.g. tmux, to allow the user to visualise the terminals for both the server and client executables. As such, the following example will assume a terminal multiplexer is being used.

1. Start the server.

# Change into the bin directory
cd bin/

# Start server executable
./server

2. Execute the client.

# Change into the bin directory
cd bin/

# Execute the client executable with appropriate inputs
./client 127.0.0.1:12345 ../images/cat2.jpg colour grey

The original and modified images for this operation can be seen below:

The original and modified images will be displayed to the user for 10 seconds, or until the user presses a key. After the display windows are removed, the modified image will be saved in the location of the original image. If an error occurs at any point, a console warning will be output to the user detailing the nature of the error.

Resetting the Images

The original images are included along with the images intended to be used by the application. To reset them, use the following commands:

1. Change into images directory.

cd images/

2. Copy original images to intended image location.

# Example using the cat2.jpg image
cp originalCat2.jpg cat2.jpg

Disclaimer

While this application features preprocessing directives to allow operation on Windows, it was built and tested on Arch Linux, and was therefore designed with POSIX-compliant systems in mind. It is possible that this application might not work on Windows, in its current state.

Footnotes

1. OpenCV adheres to the historical BGR colour space standard so all implemented filters are built around that. ↩