Enhanced image quality evaluation based on SIFT feature

Author

Guang-Lei Wen ; Gang Liu ; Si-Guo Zheng ; Shang-Kun Ning

Author_Institution

Coll. of Autom. Eng., Shanghai Univ. of Electr. Power, Shanghai, China

Volume

1

fYear

2014

fDate

13-16 July 2014

Firstpage

221

Lastpage

226

Abstract

In the study of Retinex image enhancement methods, there is a need to determine the quality of the image enhancement. But the images obtained through different Retinex image enhancement methods were quite similar. Thus, subjective evaluation criteria are not reliable. In this paper, the SIFT feature points are presented as an objective evaluation of images. Image quality evaluation method based on SIFT features were turned out to be feasible. Experimental results are consistent with the conclusions drawn by synthesizing information entropy, standard deviation, and the mean squared error.

Keywords

Gaussian processes; image enhancement; transforms; Retinex image enhancement methods; SIFT feature; image quality evaluation enhancement; information entropy synthesis; mean square error; objective evaluation; standard deviation; Abstracts; Entropy; Image recognition; Image restoration; Reliability; Standards; Enhanced image; Feature; Image quality evaluate; Retinex; SIFT (Scale Invariant Feature Transform);

fLanguage

English

Publisher

ieee

Conference_Titel

Machine Learning and Cybernetics (ICMLC), 2014 International Conference on

Conference_Location

Lanzhou

ISSN

2160-133X

Print_ISBN

978-1-4799-4216-9

Type

conf

DOI

10.1109/ICMLC.2014.7009120

Filename

7009120