MABC‐EPF: Video in‐painting technique with enhanced priority function and optimal patch search algorithm

Abstract

In this article, video in-painting technique with enhanced priority function and optimal patch searching algorithm is proposed. Initially, the source frames are extracted from video based on edge information, these frames are partitioned into two regions (1) target region (region to be in-painted) and (2) source region that is region to be considered as appropriate patch selection for in-painting the target region. A new priority function using confidence value, local structure multiplier, and motion vector uniformity to determine the in-paint location. The suitable patch is identified in source region, using correlation measure, like sum of squared difference method. The searching of the suitable patch in source region is optimized through modified artificial bee colony (MABC) algorithm. To capture more edge information, irregular patch filling is carried out to in-painting process. The in-painting process is more improved by optimal patch scaling for better vision. Finally, the in-painted source frames are combined to obtain the corrected video. The proposed method works better than the other art state applications.The proposed peak signal to noise ratio (PSNR) attains better 8.06%, 7.90%, 32.15%, and 13.06% compared with other previous methods.

Development of Image Inpainting for object removal using Enhanced Patch Priority and Matching Measures

“…Zhang et al [16] employed the combination of mean squared diffrerence and square of mean differences as a similarity metric to find the exemplar patch. The video inpainting using exemplar based methods are also evevated in recent years [17][18][19][20][21][22][23][24]. This study aimed to determine highest priority patch enhanced method to avoid dropping effect and new patch selection proces that yield favorable inpainting outcomes.…”

Adaptive Optimized Discriminative Learning based Image Deblurring using Deep CNN

“…Whereas MLP [9], SRCNN [10] and DCNN [11] are some of the discriminative learning approaches used for the above applications with the help of different training methods. Discriminative approaches have very fast testing speed but with the loss of flexibility, but optimization model-based approaches are generally time consuming with sophisticated estimation of priors to achieve excellent performance [10][11][12][13][14]. Consequently, considering the integration of these two approaches in an efficient manner to deblur the given degraded images is still a challenging task.…”