10. Code

We used C++ as our main programming language, although the algorithm should work fine in multiple languages. Since OpenCV supports C++, Java and Python (also MatLab), these will be the best language in our opinions. However, if you can use those algorithm and function in other graphic library, the algorithm should work perfectly in those languages as well.

In the future, we might implement Python version and connect with a web-page so that usercan use it directly without any programming background, but now it is just the code. Here is the GitHub repo:

https://github.com/Walter0697/Science-Fiction-Style-Image-Filtering

In this article, we would like to describe what each functions are doing in this script, so that user can modify the variables easier

void image_histogram(Mat src, Mat& dst, bool color, bool uniform, bool accumulate)

Generating the image histogram from the source image

src : source image that you would like to see the image histogram

dst : result image histogram image

color : whether the image histogram should have RGB three lines or in gray-scale one lime

uniform : whether the image histogram is uniform

accumulate : whether the image histogram is accumulate

void histogram_matching(Mat input, Mat ref, Mat& output)

Histogram matching between two images

input : source image for the histogram matching

ref : reference image for the image histogram

output : the result image after histogram matching

float getRatio(int index, float curve, float light, float offset, float height)

For getting the ratio between source image and the color picker

Variable details can be seen here:

https://walterpakwai.wixsite.com/sci-fi-filtering/post/temp4-color-pick-example

index : ratio for the corresponding index, from 0 to 255

void drawRatio(float curve, float light, float offset, float height, char* filename)

For drawing the ratio into an image, using the value from getRatio()

filename : the name of the image file after drawn

void fillColor(Mat src, Mat ref, Mat col, Mat& dst, float curve, float light, float offset, float height)

Filling color according to the ratio, details here:

https://walterpakwai.wixsite.com/sci-fi-filtering/post/temp4-color-pick-example

src : source image for the color you wanted to be filled

ref : source image after histogram matching, for reference on which color values

col : color picker image that generated from colorPickByMidPoint()

dst : result image

float getRatioByColor(int index, int midvalue, float darkness, float lightness)

Function to get the color picker value for colorPickByMidPoint()

index : color value for the corresponding index, from 0 to 255

midvalue : main theme color for the output color picker

darkness : determine how dark the darker value is

lightness : determine how bright the brighter value is

void colorPickByMidPoint(Mat& dst, int r, int g, int b, float darkness, float lightness)

Function to get the color picker from a main theme color

More information here:

https://walterpakwai.wixsite.com/sci-fi-filtering/post/7-generating-color-picker

dst : result image

r : red value of the main color

g : green value of the main color

b : blue value of the main color

darkness : determine how dark the darker value is

lightness : determine how bright the brighter value is

void drawRatio(int r, int g, int b, float darkness, float lightness, char* filename)

Different drawRatio function, for drawing the ratio by the color picker which you can see in

https://walterpakwai.wixsite.com/sci-fi-filtering/post/7-generating-color-picker

r : red value of the main color

g : green value of the main color

b : blue value of the main color

darkness : determine how dark the darker value is

lightness : determine how bright the brighter value is

filename : the name of the image file after drawn

Mat getAlpha(Mat markers, vector<vector<Point>> contours, int fst_index, int snd_index, float min, float max)

Getting the alpha value for different sections of the image using the distance transform method and normalization, more details here:

https://walterpakwai.wixsite.com/sci-fi-filtering/post/8-filling-colors-by-sections

markers : markers for different sections in the image from getMarkers()

contours : contours for different sections in the image from getMarkers()

fst_index : starting index of the section

snd_index : ending index of the section

min : minimum value for the function normalize()

max : maximum value for the function normalize()

void getMarkers(Mat src, Mat& markers, vector<vector<Point>>& contours, float light, int section)

Getting markers for different sections in the image using the watershed algorithm

More details here:

https://walterpakwai.wixsite.com/sci-fi-filtering/post/8-filling-colors-by-sections

src : source image

markers : output information that stores the markers

contours : output information that stores the contours

light : determine the size of the light source

section : determine the size of the section

void finalCombine(Mat ref1, Mat ref2, Mat dst1, Mat dst2, Mat markers, Mat& dst, bool noise)

Combining two colors together with their own alpha values by different sections

ref1 : first result image by fillColor()

ref2 : second result image by fillColor()

dst1 : first alpha image by getAlpha()

dst2 : second alpha image by getAlpha()

markers : markers to determine different sections from getMarkers()

dst : result image

noise : whether using the noise image or not