Dissertation > Excellent graduate degree dissertation topics show

Research on Algorithm of Unsupervised Color Transfer between Images Based on Clustering

Author: RaoNa
Tutor: YeXiuZuo
School: Zhejiang University
Course: Computer Applications
Keywords: Color Transfer lαβ Color Space Image Clustering User Interventions Swatches Texture Characteristics Extraction Colorization Greyscale Video
CLC: TP391.41
Type: Master's thesis
Year: 2005
Downloads: 284
Quote: 5
Read: Download Dissertation

Abstract


Transferring color between images is a method that borrows one image’s color characteristics from another. It can be utilized for color correction, colorization of old black and white movies or scientific illustrations, pseudocoloring greyscale images of luggage acquired by X-ray equipment at an airport. In medicine, greyscale images, such as Magnetic Resonance Imageing (MRI), X-ray and Computerized Tomography (CT) images can be enhanced with color for presentations and demonstrations. Recently color transfer technique has been extensively studied.This thesis presents a novel approach for transferring color between images using multiple source images. User can segment target image into several regions according to his requirement. Then we perform color transfer between individual region and corresponding source image. We experiment transferring color using two source images. Firstly, we segment the target image into background and foreground with simple binary way. Then perform color transfer between background and corresponding background source image, and foreground and corresponding foreground source image selected by user. This multiple source images technique can fuse various images’ color characteristics into one image. User can select various source images to enhance each region of target image. So this technique is an extended image enhancement approach.Unsupervised color transfer attempts to free user from laborious work. This approach can be conveniently applied to batch process and video. More specifically, eliminating the need for human intervention can be regarded as a first step towards real-time video colorization. This paper introduces a general technique for unsupervised color transfer between images based on clustering. The algorithm first uses unsupervised clustering algorithm to partition the target and source image data into a certain number of clusters with similar color and luminance. It creates optimized correspondence by matching texture character and luminance betweenclusters. Then select certain larger density pixels in each subset to construct corresponding swatch. The color information is transferred from the matched pixel in source swatch to the target one. When color transmission is done between the source and the target swatches, the final colors are assigned to each pixel of the target image by matching pixel to pixel in the target swatches using comparing texture characteristics. The proposed unsupervised color transfer between images algorithm can be used to transfer color not only between two color images, but also between a grey image and a color image. Because this technique can automatic perform color transfer process without user interventions, it can be conveniently applied to batch process and video, provided that composition, texture and luminance are sufficiently distinct.At last, this thesis analyses the technique of colorization greyscale video. We realize a colorizing greyscale video algorithm using the correlation between two conjoint frames of video. Experimental results show the algorithm works quite well.

Related Dissertations

  1. Extraction and Analysis of City Green Space Information Based on RS Image,P237
  2. Study of Several Key Issues in the Visualization of OCT Data,TP391.41
  3. Design and Implementation of Color Transfer Algorithms Based on Moment Features,TP391.41
  4. Study on Color Transferring Algorithm Based on Color-texture Information and Multi-dimensional Neighborhood Features,TP391.41
  5. Color Transfer Based on Mean-Shift Fuzzy Clustering,TP391.41
  6. Research on Clustering of Natural Scene Images with Texts Based on Random Projection,TP391.41
  7. Local Color Transfer Based on Region Matching,TP391.41
  8. Research on Image Colorization and Its Application in Video,TP391.41
  9. Research on Color Restoration for Faded Images,TP391.41
  10. The Key Technology of Fuzzy Cluster Analysis of Remote Sensing Images Based on Triple I Algorithm,TP751
  11. The Depth Estimation Based on Monocular Night Vision Image,TP391.41
  12. Fusion Technology of Infrared and Visible Images Sequences,TP391.41
  13. Concealed Weapon Detection Research Based on Image Fusion,TP391.41
  14. Recognition, Retrieval and Clustering of the Flame Image in Alumina Sintering Rotary Kiln,TP391.41
  15. Study on Example-Based Color Processing,TP391.41
  16. Real-Time Image and Video Abstraction,TP391.41
  17. Brush-based non-photorealistic rendering research,TP391.41
  18. Hierarchical Fuzzy Min-Max Clustering Algorithm and Its Application on Image Clustering,TP391.4
  19. ImageCluster: An Adaptive Heuristic Image Cluster System,TP391.41
  20. GPU-based color transfer algorithm in video processing applications,TP391.41

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