The Utilization of Cloud Computing for Facial Expression Recognition using Amazon Web Services

Face-to-face communication is a form of interaction that are done by every people, but it is hard to be implemented for blind person, especially for them to recognize their interlocutors' facial expression. Hence, a device is needed for recognizing facial expression, so blind people could recognize their interlocutors' facial expression clearly and those expressions converted into non-verbal information in type of sound. This research will use embedded device designed for machine learning called AWS DeepLens, which purposed for looking the performance of the device when using Amazon Web Services (AWS) cloud computing services for AI-trained model and compare it to on-premises deep learning. The device's outputs will inform facial expression information to the user via sound. Deep learning algorithm success will be measured in confusion matrix and its average rate is 76,16%. This proposed device utilizes cloud computing system from AWS, with objectives of system robustness and real-time usage.