Search

Real-time accurate detection of three-dimensional (3D) objects is a fundamental necessity for self-driving vehicles. Most existing computer vision approaches are based on convolutional neural networks (CNNs). Although the CNN-based approaches can achieve high detection accuracy, their high energy consumption is a severe drawback. To resolve this problem, novel energy efficient approaches should be explored. Spiking neural network (SNN) is a promising candidate because it has orders-of-magnitude lower energy consumption than CNN. Unfortunately, the studying of SNN has been limited in small networks only. The application of SNN for large 3D object detection networks has remain largely open. In this paper, we integrate spiking convolutional neural network (SCNN) with temporal coding into the YOLOv2 architecture for real-time object detection. To take the advantage of spiking signals, we develop a novel data preprocessing layer that translates 3D point-cloud data into spike time data. We propose an analog circuit to implement the non-leaky integrate and fire neuron used in our SCNN, from which the energy consumption of each spike is estimated. Moreover, we present a method to calculate the network sparsity and the energy consumption of the overall network. Extensive experiments have been conducted based on the KITTI dataset, which show that the proposed network can reach competitive detection accuracy as existing approaches, yet with much lower average energy consumption. If implemented in dedicated hardware, our network could have a mean sparsity of 56.24% and extremely low total energy consumption of 0.247mJ only. Implemented in NVIDIA GTX 1080i GPU, we can achieve 35.7 fps frame rate, high enough for real-time object detection.

Spiking neural network (SNN) is interesting both theoretically and practically because of its strong bio-inspiration nature and potentially outstanding energy efficiency. Unfortunately, its development has fallen far behind the conventional deep neural network (DNN), mainly because of difficult training and lack of widely accepted hardware experiment platforms. In this paper, we show that a deep temporal-coded SNN can be trained easily and directly over the benchmark datasets CIFAR10 and ImageNet, with testing accuracy within 1% of the DNN of equivalent size and architecture. Training becomes similar to DNN thanks to the closed-form solution to the spiking waveform dynamics. Considering that SNNs should be implemented in practical neuromorphic hardwares, we train the deep SNN with weights quantized to 8, 4, 2 bits and with weights perturbed by random noise to demonstrate its robustness in practical applications. In addition, we develop a phase-domain signal processing circuit schematic to implement our spiking neuron with 90% gain of energy efficiency over existing work. This paper demonstrates that the temporal-coded deep SNN is feasible for applications with high performance and high energy efficient.

Images typically are represented as uniformly sampled data in the form of matrix of pixels/voxels. Therefore, matrix multiply-and-accumulate (MAC) forms the core of most state-of-the-art image analysis algorithms. While digital implementation of MAC has generally been the preferred approach, high power consumption is an impediment to adopting it for medical image analysis. In this work, we present a time-domain signal processing architecture which performs MAC operations with 7bit accuracy while consuming 400X lower energy than digital implementation. The proposed architecture performs analog computation using mostly digital circuits and is suitable for scaled CMOS technologies. The proposed time-domain MAC architecture is expected to play a central role in empowering the advancement of various on-chip image analysis operations.