Point operator in image processing
This article is about overview of image processing and basic principle. I am just a student that interested in image processing.
What is Image processing?
Image processing is the methods of editing your picture before you post it on your Instagram or Facebook. It using mathematics function to produced output image from the original one.
Image processing can classify into two types, point operators and neighborhood operators. Point operators we can call pixel transformation. Neighborhood operators we can call area transformation.
Point operators or pixel transformation
According to Principles of digital image processing textbook(written by Wilhelm Burger and Mark J. Burge), its classify into 6 types
- Brightness or contrast
- Applying arbitrary intensity transformations
- Quantizing images
- Global threshold
- Gamma correction
- Color transfornmation
We can modify Brightness and contrast by using this equation.
g(x) = alpha * f(x) + beta
where g(x) as an out put image and f(x) as source image pixel.
Alpha and beta are control variables often called the gain and bias parameters, where alpha should more than 0.
Point operation and histogram
We can also use the histogram to transformation the pixel of an image but in these methods can make a permanent loss of information in the image. There are many methods for modifying brightness and contrast. These methods are part of its and usually used.
- Automatic Contrast Adjustment
we can adjust the contrast of an image by the factor. Map the lowest pixel value to zero and the highest value to the max by using this equation
fac(a) = amin + (a-alow)*amax-amin/ahigh-alow
- Histogram Equalization
this method can find and apply a point operation such that the histogram of the modified image approximates a uniform distribution. I already explain this method and how to code by using OpenCV and google colab here!. - Gamma Correction
this method uses to control the intensity and brightness of the pixel value. I have to explain that pixel value is represented the amount of light falling onto a camera sensor that emitted by a monitor or a number of toners deposited by the printer. It is usually complex and almost nonlinear in the film. In the digital world, we have to apply the intensity of the image to calibrated intensity space — human perception of intensity and minimum of bits to represent intensity range. Gamma correction is base on power function. Its use to change signal from camera to the corrected signal to show on monitor or printer. In real life, many cameras have a gamma value of around 1.3 and around 2.2 in the monitor.
That’s all from my class. Thanks for reading my article. Enjoy your image processing and don’t afraid to build an interesting project. See you next time ❤ :)
