首頁 > 網路資源 > 大同大學數位論文系統

Title page for etd-0829111-234944


URN etd-0829111-234944 Statistics This thesis had been viewed 1287 times. Download 0 times.
Author Yi-Rung Huang
Author's Email Address No Public.
Department Computer Science and Enginerring
Year 2010 Semester 2
Degree Master Type of Document Master's Thesis
Language English Page Count 58
Title A Detection Scheme Using Margin Information to Determine Regions of Interest in Images
Keyword
  • ROI
  • region of interest
  • flood-fill
  • image margin
  • image margin
  • flood-fill
  • region of interest
  • ROI
  • Abstract ROI detection is essential for many image processing schemes, such as image compression, image retrieval, and digital watermarking. This thesis presents a region of interest (ROI) detection scheme utilizing the color information of image margins and the flood-fill technique. Unlike training-based schemes, the presented scheme needs not spend time building and training object-models. It removes the background of the input image with the help of the color information of the image margin, and then discovers the ROI that contains the subject matter by means of flood-fill. The scheme contains three phases: the preprocessing phase prepares the input image for the next phase; the seed point set generation phase generates some information to be used in the following ROI determination phase, which determines the ROI in the image. The experimental results show that the presented scheme can effectively find the ROIs in different kinds of images. It can deal with images with low DOF as well as those with high DOF. Moreover, it can detect subject matters not located in the center and those whose colors are similar to the background colors.
    Advisor Committee
  • Shang-Lin Hsieh - advisor
  • Chia-Ming Chang - co-chair
  • Yo-Ping Huang - co-chair
  • Files indicate not accessible
    Date of Defense 2011-06-03 Date of Submission 2011-08-30


    Browse | Search All Available ETDs