This paper addresses the application of the Perceptron, a mathematical model based on a single biological neuron, in the context of handwritten digit classification. Using the Scilab programming environment, the study investigates the effectiveness of the Perceptron as a pattern recognizer in images. The process of creating an algorithm that collects and transforms digit images into matrices to feed the Perceptron model is described, as well as the training and classification phases. The results indicate that the Perceptron, despite being a single-layer neural network and a binary classifier, is capable of achieving satisfactory results in handwritten digit classification, highlighting its potential in pattern recognition tasks.