Correlation Method Based Analysis in Stereoscopic Images

Abstract

The stereoscopic images are presented by an intrinsic co-decomposition model. Color composition is a critical part of visual applications in design, visualization and art. The color wheel is frequently used to describe fair color combinations in geometric terms, and, in digital design, to afford a user interface to visualize and use color. To figure the correlation of inter-image or intra-image, the thin subspace clustering in super-pixel level is applied. With the limits on correlation, stereoscopic images are decayed simultaneously and the reflectance components with additional details and sophisticated contrasts are gained for the edge-preserving of super-pixel and the local reflectance correlation of pixels. Researches show that the reflectance components of co-decomposition are clearer visually. Additionally, standard deviation and information entropy of reflectance components of co-decomposition are considered to validate the efficiency quantitatively of the co-decomposition. This new geometric approach is orders of magnitude more efficient than previous work and requires no numerical optimization. We demonstrate a real-time layer decomposition tool.