Abstract: Color detection in visible spectrum is a very challenging task as the color in an image is sensitive to various factors such as sensor characteristics, illumination, optics, and ethnicity. Many researchers emphasis that there is a relationship between surface colors and illumination which is closely related problem of color constancy. For skin color detection methods, color correction stage should applied first to get better classification rate. In order to map the observed color to the correct one under some canonical illumination, prior to applying skin detection technique, common and popular solutions of illumination adaptation approaches are investigated and compared like the gray world approach, the white patch, stretch, and neural network approaches. In this paper, three new approaches have been introduced to correct the color of the image that suffers from color unbalancing. The purpose of these approaches is to correct the color natural scene and also to be used to prepare the image prior to applying skin detection technique. The first introduced color correction approach is Modified Grey World (MGW) that uses a reference value calculated using the means of RGB color components. The second approach is MGWWP to use the new MGW approach followed by white patch approach. Also, some other new approaches to solve color constancy problem are proposed during experimental tests. In this paper also, a new approach named Illumination Reflection Problem Solution (IRPSS) has been proposed to correct the skin color in the image that suffers from illumination reflection problem. Some of the skin color detection techniques are investigated and tested to show their results. The proposed IRPS approach succeeded to correct color of 95% of affected pixels from illumination reflection problem.

Abstract: The dramatic increases in computer and connectivity speeds, as per Moore’s Law, has not been accompanied by matching improvement in real-life computer performance. There continues to be frustration with performance, and users appear to be unable to experience real-life benefits of the progress in computers and connectivity. Clearly, some needless and non-transparent computer processing are continuing to be barriers to better computer services. The MHz and Kbps values of computer and internet speeds create misleading expectations about what really concern the user, such as actual web-browsing speeds. Webpage viewing is found to be of the order of 1000 times slower than file download speeds. Clearly, some real-life parameters of performance, beyond MHz and Kbps values, are needed to describe performance. In spite of the advances in hardware and software, the susceptibility to viruses appears to be greater than ever. Forced hardware and software upgrades, with mostly few benefits, often cause unforeseen problems. A survey of web-browser speeds was conducted at different locations, and at local cyber-cafes. The proposed quantitative and qualitative analysis of computer performance will help decision-making on hardware and software issues, and give some predictability to the problems that may arise.