EFFECTIVE TECHNIQUES TO ENHANCE IMAGES AND VIDEOS
Mihai P. Datcu · 2014
In this habilitation thesis are presented my research activity and the main academic results obtained after obtaining my PhD, from the second half of 2009 to date. In the following are described the main directions of my research in image processing and computer vision fields: color-to-grayscale conversion, image dehazing, fusion-based enhancing techniques, preprocessing techniques for computer vision applications and enhancing underwater images. Color-to-grayscale conversion refers to mapping three dimensional color information onto a single dimension while still preserving the original appearance, contrast and finest details. We present the main contributions of our effective decolorization algorithm that preserves the appearance of the original color image. Image dehazing refers to the process of enhancing the visibility in images degraded by haze. In outdoor environments, light reflected from object surfaces is commonly scattered due to the impurities of the aerosol, or the presence of atmospheric phenomena such as fog and haze yielding images characterized by poor contrast, lower saturation and additional noise. In our work we developed an alternative approach to solving this challenging problem. Our technique is based on the remark that the distance from the observer to the scene objects is highly correlated with the contrast degradation and the fading of the object colors. Image dehazing refers to the process of enhancing the visibility in images degraded by haze. In outdoor environments, light reflected from object surfaces is commonly scattered due to the impurities of the aerosol, or the presence of atmospheric phenomena such as fog and haze yielding images characterized by poor contrast, lower saturation and additional noise. In our work we developed an alternative approach to solving this challenging problem. Our technique is based on the remark that the distance from the observer to the scene objects is highly correlated with the contrast degradation and the fading of the object colors. Preprocessing techniques for computer vision applications. Firstly, in our previous work we introduced a decolorization technique that is suitable to match images based on local feature points