Memristive Fully Convolutional Network: An Accurate Hardware Image-Segmentor in Deep Learning

Wen, S; Wei, H; Zeng, Z; Huang, T

Memristive Fully Convolutional Network: An Accurate Hardware Image-Segmentor in Deep Learning

Wen, S

Wei, H Zeng, Z Huang, T

Permalink

Publisher:: IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
Publication Type:: Journal Article
Citation:: IEEE Transactions on Emerging Topics in Computational Intelligence, 2018, 2, (5), pp. 324-334
Issue Date:: 2018-10-01

Closed Access

	Filename	Description	Size
	Memristive_Fully_Convolutional_Network_An_Accurate_Hardware_Image-Segmentor_in_Deep_Learning.pdf	Published version	5.11 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Wen, S https://orcid.org/0000-0001-8077-7001
dc.contributor.author	Wei, H
dc.contributor.author	Zeng, Z
dc.contributor.author	Huang, T
dc.date.accessioned	2022-09-04T01:16:54Z
dc.date.available	2022-09-04T01:16:54Z
dc.date.issued	2018-10-01
dc.identifier.citation	IEEE Transactions on Emerging Topics in Computational Intelligence, 2018, 2, (5), pp. 324-334
dc.identifier.issn	2471-285X
dc.identifier.issn	2471-285X
dc.identifier.uri	http://hdl.handle.net/10453/161306
dc.description.abstract	As well known, fully convolutional network (FCN) becomes the state of the art for semantic segmentation in deep learning. Currently, new hardware designs for deep learning have focused on improving the speed and parallelism of processing units. This motivates memristive solutions, in which the memory units (i.e., memristors) have computing capabilities. However, designing a memristive deep learning network is challenging, since memristors work very differently from the traditional CMOS hardware. This paper proposes a complete solution to implement memristive FCN (MFCN). Voltage selectors are firstly utilized to realize max-pooling layers with the detailed MFCN deconvolution hardware circuit by the massively parallel structure, which is effective since the deconvolution kernel and the input feature are similar in size. Then, deconvolution calculation is realized by converting the image into a column matrix and converting the deconvolution kernel into a sparse matrix. Meanwhile, the convolution realization in MFCN is also studied with the traditional sliding window method rather than the large matrix theory to overcome the shortcoming of low efficiency. Moreover, the conductance values of memristors are predetermined in Tensorflow with ex-situ training method. In other words, we train MFCN in software, then download the trained parameters to the simulink system by writing memristor. The effectiveness of the designed MFCN scheme is verified with improved accuracy over some existing machine learning methods. The proposed scheme is also adapt to LFW dataset with three-classification tasks. However, the MFCN training is time consuming as the computational burden is heavy with thousands of weight parameters with just six layers. In future, it is necessary to sparsify the weight parameters and layers of the MFCN network to speed up computing.
dc.language	English
dc.publisher	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
dc.relation	National Natural Science Foundation of China61673187
dc.relation.ispartof	IEEE Transactions on Emerging Topics in Computational Intelligence
dc.relation.isbasedon	10.1109/TETCI.2018.2829911
dc.rights	info:eu-repo/semantics/closedAccess
dc.title	Memristive Fully Convolutional Network: An Accurate Hardware Image-Segmentor in Deep Learning
dc.type	Journal Article
utslib.citation.volume	2
pubs.organisational-group	/University of Technology Sydney
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	/University of Technology Sydney/Strength - AAII - Australian Artificial Intelligence Institute
utslib.copyright.status	closed_access	*
dc.date.updated	2022-09-04T01:16:51Z
pubs.issue	5
pubs.publication-status	Published
pubs.volume	2
utslib.citation.issue	5

Abstract:

As well known, fully convolutional network (FCN) becomes the state of the art for semantic segmentation in deep learning. Currently, new hardware designs for deep learning have focused on improving the speed and parallelism of processing units. This motivates memristive solutions, in which the memory units (i.e., memristors) have computing capabilities. However, designing a memristive deep learning network is challenging, since memristors work very differently from the traditional CMOS hardware. This paper proposes a complete solution to implement memristive FCN (MFCN). Voltage selectors are firstly utilized to realize max-pooling layers with the detailed MFCN deconvolution hardware circuit by the massively parallel structure, which is effective since the deconvolution kernel and the input feature are similar in size. Then, deconvolution calculation is realized by converting the image into a column matrix and converting the deconvolution kernel into a sparse matrix. Meanwhile, the convolution realization in MFCN is also studied with the traditional sliding window method rather than the large matrix theory to overcome the shortcoming of low efficiency. Moreover, the conductance values of memristors are predetermined in Tensorflow with ex-situ training method. In other words, we train MFCN in software, then download the trained parameters to the simulink system by writing memristor. The effectiveness of the designed MFCN scheme is verified with improved accuracy over some existing machine learning methods. The proposed scheme is also adapt to LFW dataset with three-classification tasks. However, the MFCN training is time consuming as the computational burden is heavy with thousands of weight parameters with just six layers. In future, it is necessary to sparsify the weight parameters and layers of the MFCN network to speed up computing.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/161306