Correlation optimized scanning of segmented images
Richard Pracko, Jaroslav Polec · 2008
Submitted paper deals with scanning of segmented gray-scale images using two-dimensional space filling curves. A new image processing approach based on image segmentation and correlation optimization of space-filling curves is presented. The concatenation of resulting individual segments 1-dimensional representation provides higher adjacent pixel similarity than the 1-dimensional representation of the original image. The final segments concatenated 1-dimensional representation can provide improved base for applying lossless or lossy compression methods, such as the entropic coding or predictive coding respectively. The paper analyses the adjacent pixel differences in the final 1-dimensional image representations gained using the traditional and the proposed scanning method from the entropy and prediction gain point of view to indicate the applicability of the described approach.