A Study on the Validation of Histogram Equalization as a Contrast Enhancement Technique

• Main Authors: Jasni, Mohamad Zain, Ahmed, M. Mahmood, Ahmed, M. Masroor
• Format: Conference or Workshop Item
• Language: English
• Published: IEEE 2012
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://umpir.ump.edu.my/id/eprint/6184/1/mahmood1.pdf; http://umpir.ump.edu.my/id/eprint/6184/; http://dx.doi.org/10.1109/ACSAT.2012.82
• Institution: Universiti Malaysia Pahang

Our study uncovers that histogram equalization (HE)-in a striking contrast to it's claim-is not related to enhancement of contrast. To understand this view, we start with real world images which have varying degree of image quality that almost invariably require processing to improve image contrast. For this purpose, histogram equalization including its variants is a frequently relied upon technique. HE processes image by calculating pixel density of its constituent gray levels. This mathematical model, described by HE, is neither linked to contrast nor is contrast directly included in HE equations. Therefore, the study aims to find out the factual nature of transformation functions used by HE. To understand these mathematical calculations thoroughly, the paper dismantles HE into it's building blocks. These blocks are, then, critically analyzed to understand the true relationship between HE fundamentals and contrast. This analysis' determines that HE manipulates density-not contrast-which, in turn, achieves density changes but no contrast enhancement. Hence the study concludes that HE is not a valid contrast enhancement technique