ZipNet: ZFNet-level Accuracy with 48× Fewer Parameters

Arren Matthew C. Antioquia; Daniel Stanley Tan; Arnulfo Azcarraga; Wen-Huang Cheng; Kai-Lung Hua

doi:10.1109/vcip.2018.8698672

ZipNet: ZFNet-level Accuracy with 48× Fewer Parameters

Arren Matthew C. Antioquia, Daniel Stanley Tan, Arnulfo Azcarraga, Wen-Huang Cheng, Kai-Lung Hua

Research output: Chapter in Book/Report/Conference proceeding › Conference Article in proceeding › Academic › peer-review

Abstract

With the introduction of Convolutional Neural Networks, models for image classification achieve higher classification accuracy. Based on the pattern of the design of CNN architectures, increasing the number of layers equates to a higher classification accuracy, but also increases the number of parameters and model size. This negatively affects the model training time, processing time, and memory requirement. We develop ZipNet, a CNN architecture with a higher classification accuracy than ZFNet, the winner of ILSVRC 2013, but with 48.5× smaller model size and 48.7× fewer parameters. The classification accuracy of ZipNet is higher than the performance of ZFNet and SqueezeNet on all configurations of the Caltech-256 dataset with varying number of training examples.

Original language	English
Title of host publication	2018 IEEE Visual Communications and Image Processing (VCIP)
Subtitle of host publication	Proceedings
Publisher	IEEE
ISBN (Electronic)	978-1-5386-4458-4
ISBN (Print)	978-1-5386-4459-1
DOIs	https://doi.org/10.1109/vcip.2018.8698672
Publication status	Published - 2 Jul 2018
Externally published	Yes
Event	IEEE Visual Communications and Image Processing - Taichung, Taiwan, Province of China Duration: 9 Dec 2018 → 12 Dec 2018 https://ieeexplore.ieee.org/xpl/conhome/8694905/proceeding

Conference

Conference	IEEE Visual Communications and Image Processing
Abbreviated title	VCIP 2018
Country/Territory	Taiwan, Province of China
City	Taichung
Period	9/12/18 → 12/12/18
Internet address	https://ieeexplore.ieee.org/xpl/conhome/8694905/proceeding

Keywords

Convolutional Neural Networks
Deep Learning
Image Classification
Model Compression
Object Classification

Access to Document

10.1109/vcip.2018.8698672

Cite this

@inproceedings{42f49a1e946445e7a9e9e11f2ea5162f,

title = "ZipNet: ZFNet-level Accuracy with 48× Fewer Parameters",

abstract = "With the introduction of Convolutional Neural Networks, models for image classification achieve higher classification accuracy. Based on the pattern of the design of CNN architectures, increasing the number of layers equates to a higher classification accuracy, but also increases the number of parameters and model size. This negatively affects the model training time, processing time, and memory requirement. We develop ZipNet, a CNN architecture with a higher classification accuracy than ZFNet, the winner of ILSVRC 2013, but with 48.5× smaller model size and 48.7× fewer parameters. The classification accuracy of ZipNet is higher than the performance of ZFNet and SqueezeNet on all configurations of the Caltech-256 dataset with varying number of training examples.",

keywords = "Convolutional Neural Networks, Deep Learning, Image Classification, Model Compression, Object Classification",

author = "Antioquia, {Arren Matthew C.} and Tan, {Daniel Stanley} and Arnulfo Azcarraga and Wen-Huang Cheng and Kai-Lung Hua",

year = "2018",

month = jul,

day = "2",

doi = "10.1109/vcip.2018.8698672",

language = "English",

isbn = "978-1-5386-4459-1",

booktitle = "2018 IEEE Visual Communications and Image Processing (VCIP)",

publisher = "IEEE",

address = "United States",

note = "IEEE Visual Communications and Image Processing, VCIP 2018 ; Conference date: 09-12-2018 Through 12-12-2018",

url = "https://ieeexplore.ieee.org/xpl/conhome/8694905/proceeding",

}

TY - GEN

T1 - ZipNet

T2 - IEEE Visual Communications and Image Processing

AU - Antioquia, Arren Matthew C.

AU - Tan, Daniel Stanley

AU - Azcarraga, Arnulfo

AU - Cheng, Wen-Huang

AU - Hua, Kai-Lung

PY - 2018/7/2

Y1 - 2018/7/2

N2 - With the introduction of Convolutional Neural Networks, models for image classification achieve higher classification accuracy. Based on the pattern of the design of CNN architectures, increasing the number of layers equates to a higher classification accuracy, but also increases the number of parameters and model size. This negatively affects the model training time, processing time, and memory requirement. We develop ZipNet, a CNN architecture with a higher classification accuracy than ZFNet, the winner of ILSVRC 2013, but with 48.5× smaller model size and 48.7× fewer parameters. The classification accuracy of ZipNet is higher than the performance of ZFNet and SqueezeNet on all configurations of the Caltech-256 dataset with varying number of training examples.

AB - With the introduction of Convolutional Neural Networks, models for image classification achieve higher classification accuracy. Based on the pattern of the design of CNN architectures, increasing the number of layers equates to a higher classification accuracy, but also increases the number of parameters and model size. This negatively affects the model training time, processing time, and memory requirement. We develop ZipNet, a CNN architecture with a higher classification accuracy than ZFNet, the winner of ILSVRC 2013, but with 48.5× smaller model size and 48.7× fewer parameters. The classification accuracy of ZipNet is higher than the performance of ZFNet and SqueezeNet on all configurations of the Caltech-256 dataset with varying number of training examples.

KW - Convolutional Neural Networks

KW - Deep Learning

KW - Image Classification

KW - Model Compression

KW - Object Classification

U2 - 10.1109/vcip.2018.8698672

DO - 10.1109/vcip.2018.8698672

M3 - Conference Article in proceeding

SN - 978-1-5386-4459-1

BT - 2018 IEEE Visual Communications and Image Processing (VCIP)

PB - IEEE

Y2 - 9 December 2018 through 12 December 2018

ER -

ZipNet: ZFNet-level Accuracy with 48× Fewer Parameters

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this