TY - JOUR
T1 - Fuzzy-Based Histogram Partitioning for Bi-Histogram Equalisation of Low Contrast Images
AU - Khan, Mohammad Farhan
AU - Goyal, Deepali
AU - Nofal, Muaffaq M.
AU - Khan, Ekram
AU - Al-Hmouz, Rami
AU - Herrera-Viedma, Enrique
N1 - Funding Information:
This work was supported by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah, under Grant DF-374-135-1441.
Funding Information:
This work was supported by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah, under Grant DF-374-135-1441. The authors, therefore, gratefully acknowledge DSR technical and financial support.
Publisher Copyright:
© 2013 IEEE.
PY - 2020
Y1 - 2020
N2 - The conventional histogram equalisation (CHE), though being simple and widely used technique for contrast enhancement, but fails to preserve the mean brightness and natural appearance of images. Most of the improved histogram equalisation (HE) methods give better performance in terms of one or two metrics and sacrifice their performance in terms of other metrics. In this paper, a novel fuzzy based bi-HE method is proposed which equalises low contrast images optimally in terms of all considered metrics. The novelty of the proposed method lies in selection of fuzzy threshold value using level-snip technique which is then used to partition the histogram into segments. The segmented sub-histograms, like other bi-HE methods, are equalised independently and are combined together. Simulation results show that for wide-range of test images, the proposed method improves the contrast while preserving other characteristics and provides good trade-off among all the considered performance metrics.
AB - The conventional histogram equalisation (CHE), though being simple and widely used technique for contrast enhancement, but fails to preserve the mean brightness and natural appearance of images. Most of the improved histogram equalisation (HE) methods give better performance in terms of one or two metrics and sacrifice their performance in terms of other metrics. In this paper, a novel fuzzy based bi-HE method is proposed which equalises low contrast images optimally in terms of all considered metrics. The novelty of the proposed method lies in selection of fuzzy threshold value using level-snip technique which is then used to partition the histogram into segments. The segmented sub-histograms, like other bi-HE methods, are equalised independently and are combined together. Simulation results show that for wide-range of test images, the proposed method improves the contrast while preserving other characteristics and provides good trade-off among all the considered performance metrics.
KW - Contrast enhancement
KW - dynamic range
KW - fuzzy membership function
KW - histogram equalisation
KW - image transformation
KW - optimal threshold
UR - http://www.scopus.com/inward/record.url?scp=85078750987&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85078750987&partnerID=8YFLogxK
U2 - 10.1109/ACCESS.2020.2965174
DO - 10.1109/ACCESS.2020.2965174
M3 - Article
AN - SCOPUS:85078750987
SN - 2169-3536
VL - 8
SP - 11595
EP - 11614
JO - IEEE Access
JF - IEEE Access
M1 - 8954713
ER -