Advanced neighborhood operation based image zooming in

Abstract EN

Perhaps the most interesting feature of digitally encoded image data is that the image can be analytically manipulated.

An entire field of computer graphics, named image processing, has emerged as a result of this capability.

Once the data is numerically represented and stored in a manner that corresponds to the image itself, the information can be manipulated by a computer in order to compress the image, correct defects, enhance its qualities, geometrically transform it, perform measurements, detect patterns and objects, and manipulate it in many other forms.

This paper provides an overview of the types of image processing operations and introduces an advanced algorithm to enlarge an image to get a closer look.

This algorithm changes the magnification of the image and displays the new view in a new figure.

It determines the pixel's 3-by-3 neighborhood at the output image, distributes the value of the center pixel in the input matrix to the entire pixel's neighborhood corresponding to it, and filtering the enlarged image by median.

This proposed algorithm is tested with many image classes and the results show its very good ability to perform the image zooming in.

Some preprocessing and post processing is done automatically which improve this algorithm and give its advantages over the traditional zooming operations.

The programming language MATLAB is used to realize the proposed algorithm.