• DocumentCode
    3746398
  • Title

    An improved criminisi algorithm-based image repair algorithm

  • Author

    Aiju Li;Yujie Li;Wenliang Niu;Tingmei Wang

  • Author_Institution
    Beijing Union University, Beijing, China
  • fYear
    2015
  • Firstpage
    263
  • Lastpage
    267
  • Abstract
    To obtain ideal repair effects, an improved Criminisi algorithm-based image repair algorithm was proposed specific to the shortcomings of Criminisi algorithm which costs much repairing time, its feasibility and superiority were tested. First, priority calculation was improved to find optimal to-be-repaired block; then optimal matching block search strategy was improved to find optimal matching block; finally, new confidence update modes were adopted to obtain more ideal repair effects and the simulation experiment was made to test the algorithm performance. Results indicated that compared with Criminisi algorithm not only could obtain ideal image repair effects but also could sharply reduce repair time and improve image repair efficiency.
  • Keywords
    "Maintenance engineering","Algorithm design and analysis","Optimal matching","Signal processing algorithms","Search problems","Visual effects"
  • Publisher
    ieee
  • Conference_Titel
    Image and Signal Processing (CISP), 2015 8th International Congress on
  • Type

    conf

  • DOI
    10.1109/CISP.2015.7407887
  • Filename
    7407887