Evaluation of fixed images is a task with evidence that it has been performed for centuries with facts dating from the golden age of the Greek civilization, where images represented the perfect physical beauty, but I consider for thousand years ago when mankind also built the first graphical representations there also existed the analysis and quantification of pictorial expression. Evaluation of fixed images is task that depends, among other things, on the scope in which the task is developed and according to this, it has inner connotations such as the evaluation of images from the aesthetic point of view where the qualitative trend is predominant. The purpose of this thesis is to formulate a quantitative algorithmic method represented in an evaluation grid (instrument) that allows the researcher to take decisions based on a set of images related to a specific educational text in order to select the one that best enhance the transmission of the embedded message in the text such that the pair Image-Text is more efficient for this educational purpose and is used in the construction of learning object supported by virtual platforms.