Locally-Equalized Image Contrast Enhancement Using PSO-Tuned Sectorized Equalization

Contrast enhancement is a fundamental procedure in applications requiring image processing. Indeed, image enhancement contributes critically to the success of subsequent operations such as feature detection, pattern recognition and other higher-level processing tasks. Of interest among methods available for contrast enhancement is the intensity modification approach, which is based on the statistics of pixels in a given image. However, due to variations in the imaging condition and the nature of the scene being captured, it turns out that global manipulation of an image may be vulnerable to a noticeable quality degradation from distortion and noise. This chapter is devoted to the development of a local intensity equalization strategy together with mechanisms to remedy artifacts produced by the enhancement while ensuring a better image for viewing. To this end, the original image is subdivided randomly into sectors, which are equalized independently. A Gaussian weighting factor is further used to remove discontinuities along sector boundaries. To achieve simultaneously the multiple objectives of contrast enhancement and viewing distortion reduction, a suitable optimization algorithm is required to determine sector locations and the associated weighting factor. For this, a particle-swarm optimization algorithm is adopted in the proposed image enhancement method. This algorithm helps optimize the Gaussian weighting parameters for discontinuity removal and determine the local region where enhancement is applied. Following comprehensive descriptions on the methodology, this chapter presents some real-life images for illustration and verification of the effectiveness of the proposed approach.