Dissertation > Excellent graduate degree dissertation topics show

Image Abstraction Based on Visual Attention Model and Shape-Simplifying

Author: ZhangXiCheng
Tutor: XuDan
School: Yunnan University
Course: Computer Software and Theory
Keywords: Non - uniform abstraction Visual attention Visual salience Area of ??interest map Shape to simplify Mean Curvature Flow
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 23
Quote: 0
Read: Download Dissertation

Abstract


With the development of computer graphics, people are more and more ways to simulate natural scenes. In some applications, such as network video chat, people tend to want to express through a graphical image information outside photorealistic rendering graphics, such as images, figures and background lines describe, simplify, cartoons, etc., depending on the scene to express the image the main body of information. Abstract rendering techniques appear to make up for the needs of people, abstract visual information better able to attract attention, more concise representation, to become an artist, to watch the exchange of information between the bridge. Abstract rendering of the image is a photo into a non-photorealistic effect technology, thus research in this area is not only of theoretical significance, but also has the practical value. This paper describes the abstract technology background and at this stage abstraction technology, applications, and technology status. On this basis, the focus of the paper is an the image abstract rendering of two key technologies, simplified abstract rendering algorithm that is based on abstract rendering of visual attention and shape. The first abstract drawing framework based visual attention. First, the input image is converted from the RGB color space to the CIE-Lab color space, the aim is to reduce the R, G, B three-channel correlation. The bilateral filter using an iterative construct a coherent, the edge of the significant features of the flow field, and then use the anisotropic Difference of Gaussian filter the overall line drawings extracted image. Meanwhile, based on the saliency model is derived from the input image for a visual significant function diagram, the saliency sexual function diagram can be a good indication of the human eye is concerned region. Significantly through bilateral filter, according to visual sexual function weight of non-uniformly smooth image the initial abstract rendering effects. Then visually significant function of significant FIG right weight have already come to the image a whole line FIG constraint, obtained the only visual region of non-uniform lines of Fig, then the lines add to the abstract effect diagram to enhance the contrast of the high contrast areas Finally, the article abstract image after the luminance quantization method further enhanced abstract effect. The second preliminary attempt simplified image shape. The idea is to use a vector having a cutting edge characterized in flow field to bind the mean curvature flow iterative simplification and the overall shape of the contracted image, with SHOCK filter protection important shape borders. The algorithm is mainly divided into three steps, first using an iterative bilateral filter to construct a coherent, marginally significant characteristics of the flow field. Pointed out that the image of the flow field edge tangent vector direction, make every weak edge pixels in the direction of the tangent vector approaches parallel. Based on the characteristics of the flow field, the bound of the mean curvature flow iteratively simplify and contraction of the overall shape of the image. Finally, the the SHOCK filter iteratively protection and enhancement of the image has the significant edge of the boundary structure. Since the mean curvature flow not only to the boundary shape, further constrained color characteristic diffusion, the formation of a certain style effect and therefore does not require additional processing steps, it is possible to obtain the ideal abstract effect.

Related Dissertations

  1. Study on the Feature Modulation and Selection Strategies for Modeling Visual Attention Mechanisms,TP391.41
  2. Target Detection Method Based on Visual Attention Mechanism,TP391.41
  3. Video Watermarking Research Approaches Based on Human Visual Attention,TP309.7
  4. Research on Robust Watermarking Technologies for Digital Images,TP309
  5. Technologies and Applications of Visual Saliency Detection for Image Datum,TP391.41
  6. Hypersurface in Euclidean space curvature flow Harnack inequalities research,O186.12
  7. Research on Optimization of Advanced Video Coding and Its Extension,TN919.81
  8. Several Geometric Quantities under Geometric Flows and Related Problems,O186.12
  9. Global Properties of Submanifolds in Product Manifolds,O186.12
  10. Computer Model Research of Visual Attention Based on Cooperative Work between Spatial and Object Attention,TP391.41
  11. Research on Computational Model and Experiments of Visual Attention Based on Space and Object,TP391.41
  12. Research on Some Approaches to Semantic Object Segmentation,TP391.41
  13. Research and Application on Computational Model of Visual Attention,TP391.41
  14. Research on Sorting, Tracking and Behavior Analysis of Video Object,TP391.41
  15. Research on IR Small Target Detection and Tracking Based on Attention Mechanism,TN215
  16. Attention -driven model based image retrieval method,TP391.41
  17. Kernel-based Object Tracking,TP242.62
  18. The sketch teaching new know,J214.3
  19. A Method for Image Salient Regions Detection Based on Layers and Dynamic Threshold,TP391.41
  20. Extraction of Salient Region in Image and Application in Content Based Image Retrieval,TP391.3

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile