Studies On Image Interpolation Based On The B-spline Scheme

In recent years, image interpolation has a variety of real world applications such as satellite/medical imaging, magnification of consumer photographs, video processing and so on. With the rapid development of the mobile Internet and multimedia applications, it is also becoming more and more important for the consumer electronics applications. The images with different size are delivered to users from different multimedia sources such as mobile phone, digital camera, and Internet. When the resolution of the received image is low, the users highly desire to magnify the image and appreciate it via the high resolution display devices. To construct and display high resolution images from low resolution images, simple conventional interpolation methods such as nearest neighbor, bilinear, and bicubic schemes, have been popularly used in the consumer applications. However, these methods neglect the orientations of edges and textures so as to blur interpolated images, especially the edges.This paper mainly studies the image interpolation algorithms on the B spline scheme. A novel method of adaptive B spline interpolation algorithm is proposed based on local gradient features. By introducing the inverse gradient weights, adaptive bilinear interpolation and bicubic interpolation are improved to estimate the edge directions of the interpolated points. Experimental results indicate that the proposed scheme can more excellently eliminate the checkerboard and ringing effects and protect the edges compared to traditional interpolation method. Therefore, the proposed method has an excellent edge adaptive ability.In addition, the interpolated points are obtained from the continuous interpolation curve constructed by the traditional cubic B spline interpolation. It does not highlight the edges of the image, although it can achieve a certain effect of amplification. The tangent of the B-spline curve reflects the change rate of the curve, which implicates the sharpness of the edges of the image. We propose a novel method for image interpolation based on prescribed tangent of cubic B-spline curve. First, a cubic B spline interpolation curve is constructed by discrete pixels. Then, a straight line is constructed by joining two points on the B spline curve with respect to two neighboring pixels. The value of the interpolation point between the two points is determined based on the tangent of the B spline curve, whose slope is equal to that of the above straight line. Finally, we use the Sobel operator to estimate the edge direction of the interpolation point. Experiments results indicate that the proposed method can greatly maintain the basic information of the original image, reduce the effects of image smoothing and edge blurring caused by the traditional cubic B spline interpolation, and improves the sharpness of the edges.