A Proposed Image Scaling Technique by Using Bezier Curve

Abstract

The process of resizing a digital image using geometrical transformation without changing the quality of image is known as image scaling or image resizing. Image processing such as digital image scaling has a wide range of applications on computer, mobile, and other digital devices. This paper proposes a digital image resizing (scaling) approach and explaining how the algorithms have been modified to meet accuracy and performance. Bezier curve have been used in previous works for processing in various fields while in this paper Bezier curve equations used to resize the digital image (scaling-up or scaling-down). The idea of using Bezier curve polynomial for image resizing comes from the interpolation feature of the points that located on the curve. The quality of the resized images based on the scaling factors and the control points used in Bezier curve. This work will be a useful resource for researchers who intend to apply image scaling to a real-world application because it provides a fast approach for image resizing. Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), and Signal to Noise Ratio (SNR) have been used to define the resolution of the reconstructed image. The measurements between the original image and the reconstruct image give acceptable results. The best results was found when the control points are even (n=1(0,1), n=3(0,1,2,3), …) the image will be scaled-down exactly to (1/2,1/3,1/4,….) width and height of the original image based on the scaling factor, and scaled-up to (x2, x3, x4, …) width and height of its original size based on the scaling factor, while when the control points are odd (n=2(0,1,2), n=4(0,1,2,3,4), …) the image will be scaled-down and scaled-up but some of image will be lost where the amount of lost will be based on the scaling factor.

Keywords

• Bezier curve, Image scaling, Scaling-Up, Scaling-Down, scaling factor.