Improving Accuracy of Peacock Identification in Deep Learning Model Using Gaussian Mixture Model and Speeded Up Robust Features

Tzu Ting Chen, Ding Chau Wang, Min Xiuang Liu, Chi Luen Fu, Lin Yi Jiang, Gwo Jiun Horng, Kawuu W. Lin, Mao Yuan Pai, Tz Heng Hsu, Yu Chuan Lin, Min Hsiung Hung, C. C. Chen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Data set is a most crucial aspect for the object recognition by the deep learning. The effect of training a deep learning model would be not good with an insufficient quantity of the data set. For example, there are two similar pictures which are captured from a video with a peacock and these two pictures are separated by one second in this video. Results show that the peacock was recognized by the model in the former picture but it failed in the latter picture due to the angle change of the peacock. In order to improve recognition effects of the model, we propose a system based on Gaussian Mixture Model and Speeded Up Robust Features to recognize peacocks in images. We also implement the prototype of this article and conduct a series of experiments to test the proposed solution. Furthermore, experimental results show the scheme did improve the accuracy of the complete training model.

Original languageEnglish
Title of host publicationIntelligent Information and Database Systems - 12th Asian Conference, ACIIDS 2020, Proceedings
EditorsNgoc Thanh Nguyen, Bogdan Trawinski, Kietikul Jearanaitanakij, Suphamit Chittayasothorn, Ali Selamat
PublisherSpringer
Pages565-574
Number of pages10
ISBN (Print)9783030419639
DOIs
Publication statusPublished - 2020
Event12th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2020 - Phuket, Thailand
Duration: 2020 Mar 232020 Mar 26

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12033 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2020
CountryThailand
CityPhuket
Period20-03-2320-03-26

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Improving Accuracy of Peacock Identification in Deep Learning Model Using Gaussian Mixture Model and Speeded Up Robust Features'. Together they form a unique fingerprint.

Cite this