Lightweight and Energy-Efficient Deep Learning Accelerator for Real-Time Object Detection on Edge Devices

Abstract

Tiny machine learning (TinyML) has become an emerging field according to the rapid growth in the area of the internet of things (IoT). However, most deep learning algorithms are too complex, require a lot of memory to store data, and consume an enormous amount of energy for calculation/data movement; therefore, the algorithms are not suitable for IoT devices such as various sensors and imaging systems. Furthermore, typical hardware accelerators cannot be embedded in these resource-constrained edge devices, and they are difficult to drive real-time inference processing as well. To perform the real-time processing on these battery-operated devices, deep learning models should be compact and hardware-optimized, and hardware accelerator designs also have to be lightweight and consume extremely low energy. Therefore, we present an optimized network model through model simplification and compression for the hardware to be implemented, and propose a hardware architecture for a lightweight and energy-efficient deep learning accelerator. The experimental results demonstrate that our optimized model successfully performs object detection, and the proposed hardware design achieves 1.25× and 4.27× smaller logic and BRAM size, respectively, and its energy consumption is approximately 10.37× lower than previous similar works with 43.95 fps as a real-time process under an operating frequency of 100 MHz on a Xilinx ZC702 FPGA.

The Role of Artificial Intelligence-Powered Imaging in Cerebrovascular Accident Detection

“…However, it is not feasible to have such embedded systems that do this in real time. Hence, it is compelling to develop ultra-low-power AI chips that can run inferences of AI algorithms at a faster rate and consume low energy [ 17 ]. Approaches to training AI models vary between supervised learning, unsupervised learning, transfer learning, and reinforcement learning.…”

Efficient Neural Networks on the Edge with FPGAs by Optimizing an Adaptive Activation Function

“…However, these network structures have even more network parameters, resulting in greater computational complexity and longer training times. This is evident from observing the resource requirements for the FPGA implementation of CNNs for image classification [ 7 , 8 ] and more lightweight CNNs are developed to be fit into edge devices [ 9 , 10 ]. In general, network structures with fewer parameters require shorter training times and are more suitable for systems with resource limitations on edge devices; however, they must still be capable of delivering the required performance.…”