Study Of Color Image Enhancement Based On Human Visual Characteristics And Scm

Human visual system is nonlinear and non-uniformity on processing the image, human response to images exist the differences of subjective visual effect and objective indicators parameter. In the field of image enhancement, many algorithms usually could not take into account these characteristics. The Spiking Cortical Model (SCM) is especially suitable for image processing, and it is inspired from neuronal pulse synchronizations in primate visual cortex. The negative time matrix of SCM conforms to Weber-Fechner-Law. It processes lighter areas coarsely and darker areas accurately, which is consistent with human visual characteristics.In this paper, a new image enhancement algorithm based on human visual characteristics and SCM is presented. The new algorithm achieved the better subjective quality and the better objective data. In this paper, the following works have been done:1. Analyzed the eye to the perception characteristic of the luminance and colour. Especially studied the application of Fechner logarithm law, Mach effect and masking characteristics in digital image processing.2. The histogram eqqualization algorithm, logarithmic transformation enhancement algorithm and Lapalasse algorithm are the classic image enhancement algorithms. We did experimental analysis and comparative discussion to these three algorithms deeply. At the same time, in this paper, we introduced many subjective and objective evaluation methods of the image enhancement.3. Proposed a color image enhancement algorithm based on SCM and human visual characteristic. We chose HSI color space which satisfies with human visual system. Hue component is kept unchanged and changed the luminance and saturation components. Processed luminance component through SCM, and using Mach effect to enhance the edge of image when initialization the threshold of the SCM. We used receptive field properties determining connection weighting coefficient matrix value. Processed saturation component through power stretch. We compared four enhancement algorithms using the subjective and objective evaluation methods. Experiments show that, the image which is processed through our algorithm, color clear, rich in detail and superior to other algorithms.