Automatic Image Restoration Technique For Under/Over Exposed Region
Loading...
Date
2011-07
Authors
Zakaria, Mohd Firdaus
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
Many methods have been proposed to improve the clarity, quality and to compensate the underexposed
or over-exposed regions of an image. In image restoration, the main goal is to improve the clarity
in the image. Images with high clarity level contain a lot of information. The clarity is very easily affected by
lighting, weather or equipments that have been used to capture the image. Some of these conditions such
as inconsistent lighting lead to under-exposed or over-exposed conditions where the image may suffer
information loss. Thus, this thesis puts forward an improved restoration method in order to retain the information
in the image. This proposed method only performs image restoration on the under-exposed and
over-exposed regions without degrading the correctly exposed ones. The tone mapping functions used
in this method is inspired by a shunting characteristic of the centre surround cells of the spatial filtering
output. This function is modulated differently for every pixel, according to its surround, thus minimizes the
halo effect. It also has automatic functions to recognize the under-exposed or over-exposed regions, in
which suitable correction is applied to restore these particular regions. The empirical results from series of
experiments demonstrate that the proposed method successfully improves the image quality and clarity.
The proposed method gives good results for entropy value and increases the detection rate in face detection
system. This algorithm only requires around 0.43 seconds to process an image of size 704×576
using conventional personal computer.
Description
Keywords
Image Restoration Technique , Under/Over Exposed Region