Pushing the AI envelope: merging deep networks to accelerate edge artificial intelligence in consumer electronics devices and systems

Deep neural networks (DNNs) are widely used by both academic and industry researchers to solve many long-standing problems in machine learning. There has been such a growth of research in this field, and it has been applied to so many varying problems, that it would be accurate to say that we may be living through the precursor of the singularity [1]. But regardless of one's views on artificial intelligence (AI), there is no doubt that there is a wealth of recent research that leverages the use of various DNNs to solve a broad range of pattern recognition and classification problems. Examples range from the introduction of smart speakers with intelligent assistants to the application of DNNs to solve recalcitrant problems in computer vision for autonomous vehicles. Many of these problems can have very useful applications in the design of smarter consumer electronics (CE) systems and devices. The question for CE engineers is how to leverage this wealth of academic and industry research efforts, turning them into practical DNN solutions suitable for deployment in practical devices and electronic systems.This research was funded under the Science Foundation Ireland (SFI) Strategic Partnership Program by SFI and FotoNation Ltd., project 13/SPP/I2868 on “Next Generation Imaging for Smartphone and Embedded Platforms.” We gratefully acknowledge the support of NVIDIA Corp. with the donation of a Titan X GPU used for this research. Portions of the research in this article use the CASIA-IrisV4 collected by the Chinese Academy of Sciences’ Institute of Automation.peer-reviewe