National Conference on Electronics, Signals and Communication |
Foundation of Computer Science USA |
NCESC2017 - Number 4 |
July 2018 |
Authors: K. Venkatesh Sharma, B. Rakesh |
0b8c07be-f8af-410c-8d11-217d34b5b4fb |
K. Venkatesh Sharma, B. Rakesh . Human Object Video in Painting using Shape Context Descriptor. National Conference on Electronics, Signals and Communication. NCESC2017, 4 (July 2018), 25-27.
Video inpainting is a growing area of research. In this paper we are discussing a video inpainting technique which will be useful in cases where a set of frames are missing or damaged. This technique mainly involves three steps: creating new frames, graphical model construction, estimating the correct frame. Here we limit the distance of the motion trajectory where the motion is not continuous and hence we make sure that the motion continuity is efficient. Shape context descriptor and isomap algorithm is used for finding the matching frame and clustering of the frames in the graphical model.