Efficient Edge-Based Image Interpolation Method Using Neighboring Slope Information

Sajid Khan,Dong-Ho Lee,Ghulam Mujtaba,Muhammad Asif Khan,Abdul Rehman Gilal

doi:10.1109/access.2019.2942004

Sajid Khan, Dong-Ho Lee + Show 3 more

Open Access

https://doi.org/10.1109/access.2019.2942004

Copy DOI

Abstract

This paper presents an image interpolation method that provides superior performance with low complexity. Applying a simple linear method achieves time-efficient interpolation, but often produces artifacts; however, applying other complex methods reduce unwanted artifacts at the cost of high computation time. The proposed interpolation scheme is based on the improvement of an algorithm called “Edge Slope Tracing (EST)” that predicts the slope based on the information of the adjacent slopes. Predicted slopes are used in slope-based line averaging to perform the interpolation. Slope-based line averaging is followed by post-processing of an interpolated image using two-way interpolation and thin edge correction to avoid the production of unwanted artifacts at the cost of low complexity. The proposed algorithm is very simple and simulation results indicate that it provides better or nearly equal results compared to conventional state-of-the-art algorithms and that it also reduces complexity to a great extent compared to conventional methods.

Highlights

Images and videos exist in digital formats that are discrete representations of realistic scenarios
Video interpolation is needed for viewing low resolution video on high resolution TVs, as the TV industry has evolved from analog to digital, SDTV followed by HDTV and UDTV
A new technique, Edge Slope Tracing (EEST), is introduced and used for interpolation that predicts the present slope on the basis of the information of previous slopes

Summary

INTRODUCTION

Images and videos exist in digital formats that are discrete representations of realistic scenarios. S. Khan et al.: Efficient Edge-Based Image Interpolation Method Using Neighboring Slope Information. Yang et al [11] proposed dictionary learning based interpolation method Super Resolution via Sparse Representation (SRSR) that provides pleasant results in upscalling images. NEDI is a covariance-based adaptive interpolation technique that approximates the covariance of a high-resolution image on the basis of its correspondence with low-resolution covariance. The proposed method provides better slope calculation and reset criteria It uses efficient artifact reduction and correction techniques consisting of two-way interpolation, and thin line correction. The proposed algorithm is an image interpolation method with high performance and low complexity compared to other state-of-the-art algorithms such as NEDI, ICBI and SL-ASR. The proposed algorithm is based on a simple, efficient technique, EEST, which predicts the present slope change by using the previous slope information.

SLOPE-BASED INTERPOLATION

PERFORMANCE EVALUATION

SUBJECTIVE ANALYSIS

CONCLUSION