Adaptive local exposure based region determination for non-uniform illumination and low contrast images

Abstract

Non-uniform illumination and low contrast is an issue major issue for critical and sensitive images. Recently, researchers have shown an interest in solving this issue by proposing local contrast image enhancement methods to improve the contrast of certain regions of the non-uniform illumination and low contrast images. However, most techniques concentrate on developing a specific algorithm to separately enhance only two main regions of the image such as over-exposed and under-exposed regions. Those techniques faced several issues i.e., the pixels are wrongly classified and thus make the enhancement inefficient to solve non-uniform illumination issues as well as poor contrast. These techniques are not robust, and they are specifically designed to solve a specific problem at one time. Also, these techniques have the limitation of measuring the region determination accuracy. These limitations have motivated this study to propose a new technique to solve the above-mentioned problems. An Adaptive Local Exposure Based Region Determination (ALEBRD) method is proposed to determine the image into three specific regions based on the contrast distribution, namely under-exposed, over-exposed, and well-exposed regions. The results of the proposed ALEBRD method produced better region determination performance than the other state-of-the-art methods. This research aims to determine and separate each region individually as a pre-processing stage. The prominent result from this study can be essential for the enhancement process also it can be extended to be applied in further processing such as segmentation and feature extraction. This research also proposed a new novel method to measure the region determination accuracy named Region Determination Accuracy Measurement System (RDAMS).

Keywords

• Image processing
• Non-unifrom illumination
• Region determination
• Over-exposed
• Under-exposed
• Well-exposed