Adaptive variable order polynomial based digital zoom of images

Abstract

Digital zoom is widely used in daily life from zooming a captured image to navigating through live maps. The applications are simple but application domains benefit large number of users. This manuscript proposes a novel approach to escalate zoom limits by modifying the zoom tolerance bound of an image. The approach focuses on preserving original information and transfer it to zoomed image. Digital zoom is achieved by first representing the original image as a set mathematical model representing underlying statistical parameters. The sets in model are further analyzed to calculate set variance; which in turn is used for localizing fluctuations and generate polynomial for each set. These polynomial are then used for implementing variable order interpolation scheme. Experimental results confirm that the present technique outperforms existing techniques in terms of image quality measurement parameters. The discussed approach can be implemented on RGB or grayscale images equivalently.