Enhancement on Edge and Reduction of Noise in Low Light Videos Using Discrete Wavelet Transform with JND Model

Authors

  • Bhagya H.K. , Dr. Keshaveni N

Abstract

The low contrast videos captured by camera phone are not usually convincing due to the
inappropriate illumination. The captivated videos under low light condition are of low contrast as well as
extremely unpleasant by noise. Here both the contrast enhancement and noise reduction in low light videos
using wavelet coefficients is efficiently carried out. At first, normalization is functioned on the luminance
component taken from the frame to acquire its dynamic range. Then, the frame is decomposed into low-pass
and high-pass sub-bands using Discrete Wavelet Transform (DWT). The just noticeable difference (JND)
based Contrast enhancement is obtained in the LL band and Edge enhancement and De-noising in high-pass
sub-bands HL, LH, HH. At last colour correction is attained in order to satisfy the colour distortions
rendered by contrast improvement. From the results we proved that an algorithm attains contrast
improvement, noise deduction for the dim light videos, which not only enhances the background luminance
capable for human visual recognition but also improves the contrast of the video frame.

Published

2020-11-01

Issue

Section

Articles