<![CDATA[Technology is currently developing rapidly, especially in artificial intelligence technology or AI. One part of artificial intelligence is deep learning. Deep learning is very powerful and capable of solving various problems with big data such as images, text and sound so that deep learning is widely used in research and industry needs. Face recognition and verification is an example of the application of artificial intelligence in the field of computer vision. In making facial recognition and verification using deep learning with the transfer learning method or pretrained models. Basically, the architecture of the model can be made by yourself (scratch). However, in making a model architecture that is made by yourself, it takes a lot of time to find a suitable architecture in solving certain problems, the model that is made by yourself also does not necessarily get a better level of accuracy and the inference process also takes a long time and requires sufficient data. so that the computer can recognize the image properly. The model architecture to be used is MobileNetV1 for face recognition and Facenet for face verification. Tests were carried out using a Raspberry PI with a camera-to-face distance of 40 cm, 70 cm and 100 cm and the closest distance obtained had high accuracy and an average accuracy rate of 83% with a response time of 1.26 seconds. Keywords : Transfer Learning, Raspberry PI, MobileNetV1, dan Facenet.]]>
Copyrights © 2023
