oalogo2  

AUTHOR(S): 

Woo-Sup Han, Il Song Han

 

TITLE

Neuromorphic Convolutional Recurrent Neural Network for Road Safety or Safety Near the Road

pdf PDF

 

ABSTRACT

Neuromorphic visual processing inspired by the biological vision system of brain offers an intelligent vision system of new machine vision in everyday environment. With the growing interest for detecting moving objects on the road or roadside for enhancing the safety and security, the proposed neuromorphic visual processing was tested on vehicle’s blind spot cyclist or ramming vehicle terror attack. The neuromorphic convolutional-recurrent neural network has been proposed to detect target objects and demonstrated successfully, based on the saliency of neuromorphic visual processing without complex optimization of template matching. Neuromorphic features were processed by Autoencoder and simplified Gabor filter, and detected either blind spot cyclist on the bridge or ramming vehicle in CCTV video footage of terror attack. The consistent performance of either Gabor-like filters or the small Autoencoder filters demonstrated the feasibility of real-time and robust neuromorphic vision implemented by the small embedded system.

 

KEYWORDS

Neuromorphic visual processing, Deep neural networks, Vision, Cyclist detection, Car ramming detection, Crowd movement detection

 

REFERENCES

[1] D. H. Hubel and T. N. Wiesel, Receptive fields of single neurones in the cat’s striate cortex, J. Physiol., 148, 1959, pp 574-591.

[2] Y. LeCun et al, “Deep learning,” Nature, 521, 2015, pp. 436-444.

[3] W. Han and I. S. Han, All Weather Human Detection Using Neuromorphic Visual Processing, Intelligent Systems for Science and Information, Springer International Publishing, 2014, pp25-44.

[4] M. K. Lim et al, Detection of salient regions in crowd scenes, Electronics Letters, 50, 2014, pp. 1860-1867.

[5] W. Han and I. S. Han, “Bio-Inspired Neuromorphic Visual Processing with Neural Networks for Cyclist Detection in Vehicle’s Blind Spot and Segmentation in Medical CT Images,” Proc. SAI Computing Conference 2017, pp744-750.

[6] W.J. Han et al, “Bio-inspired visual information processing- the neuromorphic approach,” WSEAS Transactions on Circuits and Systems, 9, 2010, pp. 441-452

Cite this paper

Woo-Sup Han, Il Song Han. (2017) Neuromorphic Convolutional Recurrent Neural Network for Road Safety or Safety Near the Road. International Journal of Circuits and Electronics, 2, 74-78

 

cc.png
Copyright © 2017 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0