Matrix Image Manipulation

This project wishes to provide the programmatic implementation for a paper that provides the mathematical basis of common operations in computer vision. It provides a CLI, written in Golang, that allows for basic image manipulation operations.

Running the project

Installing

Precompiled Binaries

Download the .zip corresponding to your operating system from the Releases page and extract to a directory of choice. Open the directory and run the executable as normal in your operating system.

Building from source

These instructions are OS-agnostic, for OS-specific queries please refer to the Go documentation

If not installed, download Go and Git, open a terminal window and run the following commands in the directory you would like to download the project to and run the following in order.

git clone https://github.com/MM-coder/matrix-image-manipulation.git
cd matrix-image-manipulation
go build

This will create the matrix-image-manipulation executable, in this case a .exe, to run it, you can call it in your terminal or double-click the executable

.\matrix-image-manipulation.exe

This will bring up the CLI.

Usage Guide

The CLI is currently in Portuguese, it will first ask you for a file path, for this guide we will use the provided gnome.png file that's located in the repository root. This file path must be a valid PNG file.

Qual o path do ficheiro: gnome.png

You must next pick the operation you'd like to do on the image:

Escolha uma operação:
1: Filtro Gaussiano
2: Converter para Grayscale
3: Alterar Contraste
4: Alterar Luminosidade
Escolha (1, 2, 3 ou 4):

Gaussian Filter

This will blur the image with a Gaussian filter with σ = 2.5 and a kernel size of 7 and save the output to {filename}_new.png.

Greyscale

This will convert the image to greyscale and save the output to {filename}_new.png.

Contrast

This will ask you for the value of m and b in the equation G(u) = mu+b and save the output to {filename}_new.png.

Qual o path do ficheiro: gnome.png
Escolha uma operação:
1: Filtro Gaussiano
2: Converter para Grayscale
3: Alterar Contraste
4: Alterar Luminosidade
Escolha (1, 2, 3 ou 4): 3
3.1: Insira o valor de m: 1
3.2: Insira o valor de b: -1

Tip

If you insert m=-1 and b=1 you will invert the image

Brightness

This will ask you for the value of b in the equation G(u) = u+b and save the output to {filename}_new.png.

Qual o path do ficheiro: gnome.png
Escolha uma operação:
1: Filtro Gaussiano
2: Converter para Grayscale
3: Alterar Contraste
4: Alterar Luminosidade
Escolha (1, 2, 3 ou 4): 4
4.1: Insira o valor de b:

TIP: You'll only notice a difference for larger values of b like b=50