A Double Random Matrix Design Model for Fractal Art Patterns Based on Visual Characteristics

This paper adopts the method of visual characteristics double random matrix to conduct in-depth research and analysis on design fractal art patterns. For practical application needs in graphic design, a is proposed automatically extract core base pattern based content generate four-sided continuous pattern. The first uses Canny operator for edge detection analyze area main Then, it grayscale cogeneration texture features, which best splicing selected splice extracted pattern, then, achieves two methods flat row staggered has good practicality with low complexity high versatility under premise ensuring beauty generated It can assist designers patterns, improve efficiency, save costs. In this paper, we existing image segmentation methods, adopt namely quadratic tree HV segmentation, propose new local codebook selection strategy, study degree self-adaptation different images terms strategies. makes network unable train efficiently long text prediction problems. Finally, improved algorithm tested standard living libraries, experimental results show that use codebooks coding speed significantly compared fixed fractal. face be retrieved, only necessary perform operation obtain code similarity matching, meet requirement real-time retrieval. Applying distance formula, search accuracy obtained test gallery better than histogram algorithm.