The Final Frontier

SIMD, cache, and multithreading optimizations of most common image processing procedures.

Usage

<program_name> [options] <input_file>

Options

-a, --add <value>: Perform addition operation with the specified value
-s, --sub <value>: Perform subtraction operation with the specified value
-z, --isub <value>: Perform inverse subtraction operation with the specified value
-m, --mul <value>: Perform multiplication operation with the specified value
-d, --div <value>: Perform division operation with the specified value
-c, --idiv <value>: Perform inverse division operation with the specified value
-p, --pow <value>: Perform power operation with the specified floating point value
-l, --log: Perform logarithm operation
-b, --abs: Perform absolute value operation
-n, --min <value>: Perform minimum operation with the specified value
-x, --max <value>: Perform maximum operation with the specified value
-i, --inv: Perform inversion operation
-g, --gray: Perform grayscale conversion
-e, --sobel: Perform Sobel filter operation
-f, --filter <input_file>: Perform filter operation with the specified matrix file
-k, --benchmark <value>: Perform benchmark normalized over N given iterations
-o, --output <output_file>: Specify the output image name
-h, --help: Print usage

Notes

• The image input file needs to be in the folder InputImg
• The output image files will be placed in the OutputImg folder, with two copies, one generated by the referential and one generated by the optimized functions
• The input matrix file needs to be placed in the Kernel folder (using the same format as the examples)
• Replace <program_name> with the actual name of the program you are using.
• Replace <input_file> with the path to the input file you want to process.
• Replace with the desired numerical value for the corresponding operation.
• Replace <output_file> with the desired name for the output image.

Examples

Addition operation:

<program_name> -a 5 input_file.jpg

Grayscale conversion:

<program_name> -g input_file.jpg

Sobel filter and logarithm performed one after another:

<program_name> -e -l input_file.jpg

Perform benchmark normalized over 10 iterations:

<program_name> -k 10 input_file.jpg

Perform filter operation with a matrix file:

<program_name> -f matrix.txt input_file.jpg

Help:

<program_name> -h