Comparison of Digital Image Inpainting Techniques Based on SSIM

Abstract

Digital Image Inpainting is a technique used to repair a damaged region of an image in such a way that it appears natural and continuous to an observer who is seeing it for the first time. Image inpainting is an important research area in the field of image processing used for image editing and image stitching. In this paper we compare two Digital Image Inpainting Techniques by using (structural similarity) SSIM as a metric for comparison. In the first method (Bilateral Image Inpainting. algorithm) the region to be inpainted is convolved with a bilateral kernel obtained by multiplying range and space kernels. The damaged pixel value is replaced by a weighted average of its neighbors in both space and range. A bilateral filter is an edge-preserving smoothing filter. In the second method (gradient based convolution algorithm) the damaged region is reconstructed by convolving the image with an adaptive kernel calculated by gradient of known pixels in the neighborhood of a missed pixel. These techniques are evaluated by writing a MATLAB code and using parameters like PSNR, time required for filling, SSIM, number of damaged pixels and number of reconstructed pixels.

Authors and Affiliations

Mrs Anupama Sanjay Awati, Mrs. Meenakshi R. Patil

Keywords

Related Articles

Design of Deterministic Finite Automata using Pattern Matching Strategy

An automaton is an entity work for specific computation and designed for that computation. Based on the problem automaton can be designed to satisfy the given problem. Automata have finite and infinite types. Finite aut...

slugWeighted Association Rule Mining:A Survey

Association rule mining helps to extract large transaction databases for association rule. Without taking the weight of items into account, Classical Association Rule Mining (ARM) concludes that all i...

A Fault Tolerant Priority-Based Scheme for Channel Allocation in Cellular Wireless Networks using Quorum

The effective use of channels is crucial for the performance of cellular wireless networks. The issue of fault tolerance is significant in cellular wireless networks because of mobility, poor quality of wireless links a...

Sleeping Driver Detection along with Safety Measures

Our main motive is to make a project based on a system that will help driver to stay alert on the road as well as drive or handle the vehicle only in the conscious state i.e. will be able to driver only when the person...

Data Mining for Real Time Applications and Security Issues

Data mining is used to extract knowledge from huge amount of the data Today, Data mining helps different organizations focus on customer’s behavior patterns. The research scope of data mining extended in various fields....

Download PDF file
  • EP ID EP19252
  • DOI -
  • Views 314
  • Downloads 6

How To Cite

Mrs Anupama Sanjay Awati, Mrs. Meenakshi R. Patil (2014). Comparison of Digital Image Inpainting Techniques Based on SSIM. International Journal for Research in Applied Science and Engineering Technology (IJRASET), 2(12), -. https://europub.co.uk/articles/-A-19252