Security is an essential factor in building modern systems, and user authentication is one of the most important one in this context. For such authentication, stand out today the ones based on the use of biometric sensors (e.g, fingerprint readers, face mapping) in devices such as cell phones and laptops stand out. However, this type of authentication requires specific equipments that has a cost and is not always accessible to everyone. Therefore, the general objective of this work is to authenticate people without using those specific hardwares. For this, only common microphone and cameras are used to capture audio and voice, respectively, and then methods based on Machine Learning ( textit Machine Learning - ML) are used to identify a previously registered user. To evaluate this work, a graphic tool for registering and authenticating people was developed, which was used internally by members of the laboratory.