Parallel Image Processing for Coursera Projects

Overview

This uses CUDA for converting multiple images into grayscale image.

All images in a given folder will be converted and
saved input the given output folder.

The number of parallel CUDA streams can be specified.

Usage

Command line arguments

Usage: image_processing.exe [OPTIONS] --inputs <INPUTS> --outputs <OUTPUTS>

Options:
  -i, --inputs <INPUTS>    input file folder
  -o, --outputs <OUTPUTS>  output file folder
  -s, --streams <STREAMS>  Number of streams [default: 1]
  -h, --help               Print help
  -V, --version            Print version

Example

run.bat run.sh

Will load all images in folder images and will convert each into new folder out

Before Run:

Run:

After Run:

Build and execute on Windows

1. build.bat
2. run.bat

Build and execute on Linux

1. build.sh
2. run.sh

Code Organization

README.md

this text you are reading

build.bat build_kernel.bat run.bat

build and execute scripts for Windows

build.sh build_kernel.sh run.sh

build and execute scripts for Linux

images

folder with input images

src/main.rs

application in rust

kernels/filter.cu

CUDA kernel for grayscale conversion