Discipline: Computer Sciences and Information Management
Subcategory: Computer Science & Information Systems
Session: 2
Room: Exhibit Hall A
Malcolm Phipps - Prairie View A&M University
Deep learning is an artificial intelligence function that imitates the workings of a human brain in processing data and decision making. As technology keeps improving Artificial Intelligence will be the future of this world. At the meantime, with the development of IoT technology, it is becoming more and more critical to deploy machine learning technology into embedded systems. The problem is the lack of resources in embedded systems, makes it difficult to perform machine learning tasks in a real time manner. Based on this observation, an embedded system is designed and built to support real time facial recognition, object identification, and conversation. This system implements deep learning in opencv to provide the ability for real time facial recognition and object detection and IBM Watson conversational system for human-machine conversation. This is done on a Raspberry Pi 3 with help from the Intel Neural Compute Stick 2 (NCS2) to speed up the machine learning process. Intel NCS2 significantly increases the response speed to ensure that the real time reaction of the system replicates that of a human brain. This system will be the major part of the perception system for our future autonomous robotics system. Also, it will be used together with our near infrared spectroscopy for our future human brain and mental researcH.
Funder Acknowledgement(s): NSF HBCU-UP program
Faculty Advisor: Yonghui Wang, yowang@pvamu.edu
Role: Used deep learning with embedded systems