Convolutional neural networks (CNNs) have become the dominant neural network architecture for solving many state-of-the-art (SOA) visual processing tasks. Even though graphical processing units are most often used in training and deploying CNNs, their power efficiency is less than 10 GOp/s/W for single-frame runtime inference. We propose a flexible and efficient CNN accelerator architecture called NullHop that implements SOA CNNs useful for low-power and low-latency application scenarios. NullHop exploits the sparsity of neuron activations in CNNs to accelerate the computation and reduce memory requirements. The flexible architecture allows high utilization of available computing resources across kernel sizes ranging from 1x1 to 7x7. NullHop can process up to 128 input and 128 output feature maps per layer in a single pass. We implemented the proposed architecture on a Xilinx Zynq field-programmable gate array (FPGA) platform and presented the results showing how our implementation reduces external memory transfers and compute time in five different CNNs ranging from small ones up to the widely known large VGG16 and VGG19 CNNs. Postsynthesis simulations using Mentor Modelsim in a 28-nm process with a clock frequency of 500 MHz show that the VGG19 network achieves over 450 GOp/s. By exploiting sparsity, NullHop achieves an efficiency of 368%, maintains over 98% utilization of the multiply-accumulate units, and achieves a power efficiency of over 3 TOp/s/W in a core area of 6.3 mm₂. As further proof of NullHop's usability, we interfaced its FPGA implementation with a neuromorphic event camera for real-time interactive demonstrations.

The problem of finding stereo correspondences in binocular vision is solved effortlessly in nature and yet it is still a critical bottleneck for artificial machine vision systems. As temporal information is a crucial feature in this process, the advent of event-based vision sensors and dedicated event-based processors promises to offer an effective approach to solving the stereo matching problem. Indeed, event-based neuromorphic hardware provides an optimal substrate for fast, asynchronous computation, that can make explicit use of precise temporal coincidences. However, although several biologically-inspired solutions have already been proposed, the performance benefits of combining event-based sensing with asynchronous and parallel computation are yet to be explored. Here we present a hardware spike-based stereo-vision system that leverages the advantages of brain-inspired neuromorphic computing by interfacing two event-based vision sensors to an event-based mixed-signal analog/digital neuromorphic processor. We describe a prototype interface designed to enable the emulation of a stereo-vision system on neuromorphic hardware and we quantify the stereo matching performance with two datasets. Our results provide a path toward the realization of low-latency, end-to-end event-based, neuromorphic architectures for stereo vision.

Edge artificial intelligence hardware targets mainly inference networks that have been pretrained on massive datasets. The field of few-shot learning looks for methods that allow a network to produce high accuracy even when only a few samples of each class are available. Siamese networks can be used to tackle few-shot learning problems and are unique because they do not require retraining on the new samples of the new classes. Therefore they are suitable for edge hardware accelerators which often do not include on-chip training capabilities. This work describes improvements to a baseline Siamese network and benchmarking of the improved network on edge platforms. The modifications to the baseline network included adding multi-resolution kernels, a hybrid training process as well a different embedding similarity computation method. This network shows an average accuracy improvement of up to 22% across 4 datasets in a 5-way, 1-shot classification task. Benchmarking results using three edge computing platforms (NVIDIA Jetson Nano, Coral Edge TPU and a custom convolutional neural network accelerator) show that a Siamese classifier can run on these devices at reasonable frame rates for real-time performance, between 3 frames per second (FPS) on Jetson Nano and 60 FPS on the Edge TPU. By increasing the weight sparsity during training, the inference time of a network with 25% weight sparsity increases by 10 FPS but with only 1% drop in accuracy.

No abstract

No abstract

No abstract

No abstract