Open Access Open Access  Restricted Access Subscription Access

A Comparison on Different Image Completion Methods


Affiliations
1 Mar Baselious Christian College of Engineering and Technology, Kuttikanam, Peermade, Kerala, India
 

Image inpainting and geometric rearrangements are closely linked and can be understood simultaneously with the use of algorithms and shift map. This paper also discusses the formulation and algorithm of inpainting by correspondence map with the clear cut notations and equations. The correspondence map is defined as linking each and every blank or missing pixel to the some pixels, where that pixel’s value is in the seed image. Here the algorithm is a descent for E(F). This paper also describes a new representation of geometric rearrangement of images using operations such as graph labelling as part of image editing, where the shift map represents the selected label for each output pixel. Pixel rearrangement, pixel saliency and removal, smoothness term for pixel pairs is described in detail.
User
Notifications
Font Size

Abstract Views: 336

PDF Views: 0




  • A Comparison on Different Image Completion Methods

Abstract Views: 336  |  PDF Views: 0

Authors

Merin Varghese
Mar Baselious Christian College of Engineering and Technology, Kuttikanam, Peermade, Kerala, India
Sherin Chacko Thomas
Mar Baselious Christian College of Engineering and Technology, Kuttikanam, Peermade, Kerala, India
S. Manimurugan
Mar Baselious Christian College of Engineering and Technology, Kuttikanam, Peermade, Kerala, India

Abstract


Image inpainting and geometric rearrangements are closely linked and can be understood simultaneously with the use of algorithms and shift map. This paper also discusses the formulation and algorithm of inpainting by correspondence map with the clear cut notations and equations. The correspondence map is defined as linking each and every blank or missing pixel to the some pixels, where that pixel’s value is in the seed image. Here the algorithm is a descent for E(F). This paper also describes a new representation of geometric rearrangement of images using operations such as graph labelling as part of image editing, where the shift map represents the selected label for each output pixel. Pixel rearrangement, pixel saliency and removal, smoothness term for pixel pairs is described in detail.