PictroLib

PictroLib is a classic crossplatform image processing library developed to be integrated as a backend for an image editing application that I plan to make. It can perform the basic operations on both Black-White and coloured images and can also be extended to add more functionalitites depending on the need of the application.

Currently it supports only .BMP files and has the appropriate mechanisms to read, parse and manipulate the image data. Some features include -

Rotation/Mirror

The current implementation support the followings directions.

LEFT
RIGHT
UPSIDEDOWN
MIRROR

Brightness

The current implementation support the followings options

increaseBrightess(factor)
decreaseBrightness(factor)

Contrast

The contrast is created by equalizing the image histogram.

Add Images here

Negetive

The blur is calculated by simply inverting the pixel values at each pixel.

Blur

The blur is calculated by convolution of the image with the KERNAL_FILTER

static const float KERNAL_FILTER[3][3] =  {
    {1.0/9.0, 1.0/9.0, 1.0/9.0},
    {1.0/9.0, 1.0/9.0, 1.0/9.0},
    {1.0/9.0, 1.0/9.0, 1.0/9.0}
};

Sepia

The values for sepia filter is taken from the internet and in the same way we can add many other classic filters

r = (m_imageData[i][0] * 0.393) + (m_imageData[i][1] * 0.769) + (m_imageData[i][2] * 0.189);
g = (m_imageData[i][0] * 0.349) + (m_imageData[i][1] * 0.686) + (m_imageData[i][2] * 0.168);
b = (m_imageData[i][0] * 0.272) + (m_imageData[i][1] * 0.534) + (m_imageData[i][2] * 0.131);

Maximum

The maximum filter assigns each pixel with a value which is the maximum in a given viscinity. For 3, 5, 9 for 1, 2, 3 level deep neighbours.

applyMaxFliter(depth)

Minimum

The minimum filter assigns each pixel with a value which is the maximum in a given viscinity. For 3, 5, 9 for 1, 2, 3 level deep neighbours.

applyMinFilter(depth)

Median

The median filter assigns each pixel with a value which is the median in a given viscinity. For 3, 5, 9 for 1, 2, 3 level deep neighbours.

applyMedian(depth)

Convolution Masks/Edge Detection

We use a variety of populare edge dectedtion masks which after convolution with the image data give us some solid edges in any desired direction.

LINE_DETECTOR_HOR_MASK,
LINE_DETECTOR_VER_MASK,
LINE_DETECTOR_LDIA_MASK,
LINE_DETECTOR_RDIA_MASK,

//Prewitt Masks
PREWITT_HOR,
PREWITT_VER,

//Sobel Masks
SOBEL_HOR,
SOBEL_VER,

//Robinson Masks(Compass Operator)
ROBINSON_NORTH,
ROBINSON_NORTHEAST,
ROBINSON_EAST,
ROBINSON_SOUTHEAST,
ROBINSON_SOUTH,
ROBINSON_SOUTHWEST,
ROBINSON_WEST,
ROBINSON_NORTHWEST,

//Kirsch Masks(Compass Operator)
KIRSCH_NORTH,
KIRSCH_NORTHEAST,
KIRSCH_EAST,
KIRSCH_SOUTHEAST,
KIRSCH_SOUTH,
KIRSCH_SOUTHWEST,
KIRSCH_WEST,
KIRSCH_NORTHWEST,

//Laplacian Masks
LAPLACIAN_NEGETIVE,
LAPLACIAN_POSITIVE,

//High Pass FIlter(Sharpner)
IMAGE_SHARPNER

Gaussian Noise

Produce Gaussian noise for a given image

Salt and Pepper Noise

Produce grainy noice with a given intensity. For example in this case we generate two grainy noises on a sample image with .1 and .3 intensity.

generateNoise(intensity)

That is it for the Library, there are a few more operations for which I have not attached the photos but you can play with the convolution masks and even add some of your own in the Constants file and attach the results here. :)