Developing image/video enhancement methods and its performance evaluation using face recognition algorithms

Image enhancement has been in practice for decades. The Retinex theory is one of the best image enhancement tools. It was introduced by Edwin Land in 1986. Extensive research has been done since its introduction in the late 1990s It provides good dynamic range compression and color constancy but they are computationally expensive and cannot be applied for a broader class of images.

This thesis introduces a novel image enhancement with color restoration algorithm by combination of retinex and fast transform tools. The key advantages for the developed algorithm are a) faster due to its simple and robust enhancement technique, b) it provides better enhancement in terms of color and contrast and 3) reduced computational complexity. An image fusion technique was developed, three different algorithms were combined to produce a output image with better contrast and color image processing. Quantitative analysis of the algorithm was done on a broader class of images ranging from visual, thermal, underexposed, over exposed and bio-medical images. Different types of thermal and natural scene NASA images have been tested, along with other imagery. Furthermore, performance evaluation of image enhancement was also made by using face recognition algorithm. The image enhancement algorithm is further applied in video enhancement for color and black & white images. Additionally, a graphical user interface was developed for easy user interaction.