Picture reconstruction methods based on multilayer perceptrons

Yiheng Hu, Zhihua Hu, Jin Chen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Multilayer Perceptrons (MLPs) with Back Propagation (BP) training algorithm have been successfully utilized for solving a wide variety of real world engineering problems, such as pattern classification, character recognition, function approximation, clustering and forecasting. In this paper, a MLP classifier is built by using BP algorithm for picture reconstruction. According to the engineering consideration of the effectiveness and efficiency, the optimal parameters of neural networks are carefully selected. Moreover, the contribution of data size for reconstruction accuracy and time consuming are checked. The number of epochs for training is also an important fact for time consuming which has strong relationship with the avoid overtraining algorithm.

Original languageEnglish
Title of host publicationRecent Developments in Intelligent Systems and Interactive Applications
Subtitle of host publicationProceedings of the International Conference on Intelligent and Interactive Systems and Applications, IISA 2016
EditorsFatos Xhafa, Srikanta Patnaik, Zhengtao Yu
PublisherSpringer, Cham
Pages337-342
Number of pages6
Volume541
ISBN (Electronic)9783319495682
ISBN (Print)9783319495675
DOIs
Publication statusPublished - 30 Nov 2016
Externally publishedYes
EventInternational Conference on Intelligent and Interactive Systems and Applications - Shanghai, China
Duration: 25 Jun 201626 Jun 2016

Publication series

NameAdvances in Intelligent Systems and Computing
PublisherSpringer
Volume541
ISSN (Print)2194-5357
ISSN (Electronic)2194-5365

Conference

ConferenceInternational Conference on Intelligent and Interactive Systems and Applications
Abbreviated titleIISA2016
Country/TerritoryChina
CityShanghai
Period25/06/1626/06/16

Cite this