A Hardware System for Fast AER Object Classification with On-chip Online Learning

Huang, Jinguo; He, Wei; Zhou, Xichuan; He, Junxian; Wang, Ying; Shi, Cong; Lin, Yun

doi:10.1109/icta48799.2019.9012908

Cited by 1 publication

(5 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As described in Equations (1) and (2), the inference procedure in Random Ferns contains multiplication and division operations, which require many expensive computing resources and result in high system latency. To solve this problem, we converted Equation (1) into (3) via a logarithmic transform [ 32 ]:

…”

Section: Algorithm Reviewmentioning

confidence: 99%

“…Therefore, in this paper, we propose a high-speed low-cost VLSI hardware system dedicated to AER object classification tasks with fast on-chip online learning capabilities [ 32 ], based on such lightweight statistical algorithms utilizing the Random Ferns classifier [ 30 ]. The proposed system mainly consists of 3 modules: a motion detector, a bank of binary feature extractors, and a Random Ferns engine.…”

Section: Introductionmentioning

confidence: 99%

“…This work has partly been released in [ 32 , 33 ]. In [ 33 ], a premature version of our work without on-chip online capability was reported, where the classifier parameters of the Random Ferns engine had to be learned offline on a general-purpose computer.…”

Section: Introductionmentioning

confidence: 99%

“…In [ 33 ], a premature version of our work without on-chip online capability was reported, where the classifier parameters of the Random Ferns engine had to be learned offline on a general-purpose computer. Later on, our completed system incorporating on-chip online learning was reported in [ 32 ]. However, both Ref.…”

Section: Introductionmentioning

confidence: 99%

“…However, both Ref. [ 33 ] and [ 32 ] are brief conference reports, with neither sufficient design details nor elaborate experimental results on more DVS datasets other than MNIST-DVS. In contrast, this paper covers all of the system architecture and circuit design details with sufficient experimental measurements on more DVS datasets: Poker-DVS and Posture-DVS.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A High-Speed Low-Cost VLSI System Capable of On-Chip Online Learning for Dynamic Vision Sensor Data Classification

Huang

Wang

et al. 2020

Sensors

Self Cite

View full text Add to dashboard Cite

This paper proposes a high-speed low-cost VLSI system capable of on-chip online learning for classifying address-event representation (AER) streams from dynamic vision sensor (DVS) retina chips. The proposed system executes a lightweight statistic algorithm based on simple binary features extracted from AER streams and a Random Ferns classifier to classify these features. The proposed system’s characteristics of multi-level pipelines and parallel processing circuits achieves a high throughput up to 1 spike event per clock cycle for AER data processing. Thanks to the nature of the lightweight algorithm, our hardware system is realized in a low-cost memory-centric paradigm. In addition, the system is capable of on-chip online learning to flexibly adapt to different in-situ application scenarios. The extra overheads for on-chip learning in terms of time and resource consumption are quite low, as the training procedure of the Random Ferns is quite simple, requiring few auxiliary learning circuits. An FPGA prototype of the proposed VLSI system was implemented with 9.5~96.7% memory consumption and <11% computational and logic resources on a Xilinx Zynq-7045 chip platform. It was running at a clock frequency of 100 MHz and achieved a peak processing throughput up to 100 Meps (Mega events per second), with an estimated power consumption of 690 mW leading to a high energy efficiency of 145 Meps/W or 145 event/μJ. We tested the prototype system on MNIST-DVS, Poker-DVS, and Posture-DVS datasets, and obtained classification accuracies of 77.9%, 99.4% and 99.3%, respectively. Compared to prior works, our VLSI system achieves higher processing speeds, higher computing efficiency, comparable accuracy, and lower resource costs.

show abstract

…”

Section: Algorithm Reviewmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%