Research On Image Filter And Stylization

At present,image filtering and stylization which are regarded as a research field in digital image processing,have attracted many scholars' research and attention.With the development of society,image filtering and stylization have gradually entered into everyone's life before you know it,they can make people's life become colorful and enrich people's spiritual world at the same time.In addition to the many applications in film and television art,multimedia entertainment and other fields,image filtering and stylization also have a wide application space in medical,engineering and other fields.This paper,including filters(emboss filter,light filter,nostalgia filter,soft light filters,sketch filter),image abstraction style and image appearance transfer style and other realization method of a variety of filters and style,by learning the basic knowledge of image processing,design algorithm,repeated experiments and improvement,the ideal algorithm results are obtained finally.This paper describes processing algorithms of several improved styles.As in the section of image abstraction style,a simple and effective real time method of abstract image is proposed,meanwhile,retaining the important information of perception.In the method,image abstraction style is implemented via nonlinear diffusion filtering,edge detection and extraction,soft brightness quantization and other steps.In the section of image appearance transfer style,an image tone mapping algorithm based on bilateral filtering is proposed,which is based on a compression method named Durand.The image is divided into the basic layer and the detail layer through the bilateral filtering,and the basic layer and the template image are fused,to achieve the purpose of the input image to show the style of the template image.Image filtering and stylization which is studied in the paper,also has a good development prospects,is a notable research field.