Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Automatic neural network search method for Open Set Recognition

Abstract: 

Real-world recognition or classification tasks in computer vision are not apparent in controlled environments and often get involved in open set. Previous research work on real-world recognition problem is knowledge- and labor-intensive to pursue good performance for there are numbers of task domains. Auto Machine Learning (AutoML) approaches supply an easier way to apply advanced machine learning technologies, reduce the demand for experienced human experts and improve classification performance on close set. This paper proposes an automatic neural network search method for designing effective convolution neural network (CNN) models for open set recognition (OSR). Feature distribution information is explicitly incorporated into the main objective. So during the search process, the sampled models will enlarge inter-class differences and reduce intra-class variations. We design a flexible search space based on classic CNN models to diversify neural architectures and also add some search principles to limit the size of the search space. Experimental results on CIFAR-10 and Dunhuang historical Chinese datasets show that our approach improves performances on both close and open set. Comparing with the other two OSR algorithms, our method also achieves the best performance.

up
0 users have voted:

Paper Details

Authors:
Li Sun, Xiaoyi Yu, Liuan Wang, Jun Sun, Hiroya Inakoshi, Ken Kobayashi, Hiromichi Kobashi
Submitted On:
17 September 2019 - 4:35am
Short Link:
Type:
Poster
Event:
Presenter's Name:
Li Sun
Paper Code:
3142
Document Year:
2019
Cite

Document Files

ICIP_sunli_v3.pdf

(17)

Subscribe

[1] Li Sun, Xiaoyi Yu, Liuan Wang, Jun Sun, Hiroya Inakoshi, Ken Kobayashi, Hiromichi Kobashi, "Automatic neural network search method for Open Set Recognition", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4657. Accessed: Dec. 14, 2019.
@article{4657-19,
url = {http://sigport.org/4657},
author = {Li Sun; Xiaoyi Yu; Liuan Wang; Jun Sun; Hiroya Inakoshi; Ken Kobayashi; Hiromichi Kobashi },
publisher = {IEEE SigPort},
title = {Automatic neural network search method for Open Set Recognition},
year = {2019} }
TY - EJOUR
T1 - Automatic neural network search method for Open Set Recognition
AU - Li Sun; Xiaoyi Yu; Liuan Wang; Jun Sun; Hiroya Inakoshi; Ken Kobayashi; Hiromichi Kobashi
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4657
ER -
Li Sun, Xiaoyi Yu, Liuan Wang, Jun Sun, Hiroya Inakoshi, Ken Kobayashi, Hiromichi Kobashi. (2019). Automatic neural network search method for Open Set Recognition. IEEE SigPort. http://sigport.org/4657
Li Sun, Xiaoyi Yu, Liuan Wang, Jun Sun, Hiroya Inakoshi, Ken Kobayashi, Hiromichi Kobashi, 2019. Automatic neural network search method for Open Set Recognition. Available at: http://sigport.org/4657.
Li Sun, Xiaoyi Yu, Liuan Wang, Jun Sun, Hiroya Inakoshi, Ken Kobayashi, Hiromichi Kobashi. (2019). "Automatic neural network search method for Open Set Recognition." Web.
1. Li Sun, Xiaoyi Yu, Liuan Wang, Jun Sun, Hiroya Inakoshi, Ken Kobayashi, Hiromichi Kobashi. Automatic neural network search method for Open Set Recognition [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4657