The Mexican Sign Language (Lenguaje de Señas Mexicano, LSM) consists of movements of the human body and hands to communicate, this is used by the deaf-mute to articulate thoughts and emotions. Thus emerges the initiative to device an Automatic Computer System for the Translation of Sign Language to Text, in order to facilitate communication to deaf and non-deaf people. The system is based on pattern recognition. This recognition is applied continuously using a capture device, in this case a Web camera using the Visual Studio IDE specifically with the OpenCV library. The recognition of the signs is made by comparing a matrix, previously obtained by means of several algorithms for the extraction of the object of interest, in this case the hand. In tests for the recognition of numbers, an average reliability percentage of 80%was obtained for the patterns analyzed.