The big volume of data and information generated today, has forced the use of different means to achieve better communication.The most commonly used are text and image, which are related by their ecphrasis, which is the verbal representation of a visual representation.Therefore, this article is made in order to identify text/image relationships (TRI), and proposes a classification and visualization model based on the principle of similarity in nature, in order to avoid the proliferation of isolated or independent RTIs.This is achieved through a review and characterization of sources that allow an analysis to be made of the different RTIs, which are processed using documentary techniques and information analysis, allowing a clarification and a differentiated use.In the article, different RTI are based on reality, the message, logical semantics and relevance in comics; likewise, it evidences RTI functions, its different dialogical relationships and ends with illustrative ways.Finally, presents the 2D and 3D classification and representation models allow to identify classes, hierarchies and spatial location of the RTI structure.Such classification and representation models are key to the objective interpretation and organization of RTI and texts, not only by the reader, but also by computer systems for the extraction of characteristics in context.