Image contrast enhancement using histogram equalization: a bacteria colony optimization approach

Abstract

Histogram equalization (HE) is a popular method for image contrast enhancement that probabilistically maps the existing tonal levels of image to a new set of intensity levels. Despite simplicity, conventional global HE (GHE) has several limitations including visually disturbing false contouring, loss of original image features and wrong color representation in case of color images. Further modifications of GHE method have shown considerable improvement but many such techniques lack retention of original image characteristics. Computational intelligence algorithms can be a potential paradigm to address those limitations. This paper explores the scope of bacteria colony optimization (BCO) to obtain contrast enhancement while maintaining the original image characteristics. The fitness function has been formulated in frequency domain upon rigorous analysis in comparison with the results of existing contrast enhancement techniques. The paper includes implementation results with standard databases and the comparative evaluations. The comparative visual and objective evaluations confirm potential of BCO for improved performance.