This paper presents a method to classify images in a WEB service based on Open Source technology to increase the usability of the system, particularly to users with disabilities or "non-expert" users. In addition with this implementation, we have studied new methods to search images on Internet with a semantic approach and the relationship between the size of the screen and the quality perceived by the user as other important parameter to improve the usability of a WEB image system. The WEB service programmed will allow us to validate (or no) two main hypotheses: (HI): An iconographic system to classify images increases the usability of interfaces where we can look, index or find those images. This point is very important to approach the technology to users with disabilities, but it is necessary to consider the particularities of every disability and how the system can adapted to them. (H2): The technical data of the image and the action of visualization them, are closely connected with the quality and the emotion perceived by the user. If we can find the optimum relationship among the technical characteristics of the image and the distance or size of the screen, we can optimize the emotional information of the image that remains reflected in the user and the interaction with the system.