Quadtree-segmented image compression method using vector quantization and cubic B-spline interpolation

Abstract

A novel hybrid image compression method based on quadtree is proposed, involving two techniques: vector quantization (VQ) and cubic B-spline interpolation. Traditional block-based image compression methods, such as VQ, block truncation coding and others, do not take the relationship of neighbouring blocks into consideration, which results in a limit to their compression rate. To put it another way, if the relationships between neighbouring blocks can be included in the scheme by application of the quadtree technique and the VQ method, as well as cubic B-spline interpolation, which is exactly what is attempted in this paper, the performance of image compression can be improved significantly. According to the experimental results, this method can achieve satisfactory results in terms of the compression rate, while maintaining an acceptable image quality compared with block-based image compression methods.