Genişletilmiş deeplabv3 mimarisi ile anlamsal bölütleme

Salih Can Yurtkulu; Yusuf Huseyin Sahin; Gozde Unal

doi:10.1109/SIU.2019.8806244

Genişletilmiş deeplabv3 mimarisi ile anlamsal bölütleme

Salih Can Yurtkulu, Yusuf Huseyin Sahin, Gozde Unal

Istanbul Technical University

Araştırma sonucu: Kitap/Rapor/Konferans Bildirisinde Bölüm › Konferans katkısı › bilirkişi

123 Atıf (Scopus)

Özet

In this work, semantic segmentation has been dealt with convolutional neural networks (CNN) which is a widely used recent approach in the field of computer vision. In the experiments using Cityscapes dataset, the images are scaled by various rates and the CNN architecture named DeepLabv3 is trained with different hyperparameters using these images. After the training phase, the success rates of the trained models were compared. The most successful DeepLabv3 model has achieved a success rate of 78.83% on Cityscapes test set. Afterwards, an ensemble of two different DeepLabv3 models and the Extended DeepLabv3 model is tested. In test results, while the success rate remains nearly the same, an increase in classes such as road and sidewalk is observed.

Tercüme edilen katkı başlığı	Semantic segmentation with extended DeepLabv3 architecture
Orijinal dil	Türkçe
Ana bilgisayar yayını başlığı	27th Signal Processing and Communications Applications Conference, SIU 2019
Yayınlayan	Institute of Electrical and Electronics Engineers Inc.
ISBN (Elektronik)	9781728119045
DOI'lar	https://doi.org/10.1109/SIU.2019.8806244
Yayın durumu	Yayınlandı - Nis 2019
Etkinlik	27th Signal Processing and Communications Applications Conference, SIU 2019 - Sivas, Turkey Süre: 24 Nis 2019 → 26 Nis 2019

Yayın serisi

Adı	27th Signal Processing and Communications Applications Conference, SIU 2019

???event.eventtypes.event.conference???

???event.eventtypes.event.conference???	27th Signal Processing and Communications Applications Conference, SIU 2019
Ülke/Bölge	Turkey
Şehir	Sivas
Periyot	24/04/19 → 26/04/19

Bibliyografik not

Publisher Copyright:
© 2019 IEEE.

Keywords

Convolutional neural networks (CNN)
Deep learning
Semantic segmentation

Belgeye Erişim

10.1109/SIU.2019.8806244

Diğer Dosyalar ve Linkler

Link to publication in Scopus

Alıntı Yap

Yurtkulu, S. C., Sahin, Y. H., & Unal, G. (2019). Genişletilmiş deeplabv3 mimarisi ile anlamsal bölütleme. In 27th Signal Processing and Communications Applications Conference, SIU 2019 Makale 8806244 (27th Signal Processing and Communications Applications Conference, SIU 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SIU.2019.8806244

@inproceedings{0f9c956691384794a82dc14c65e63989,

title = "Geni{\c s}letilmi{\c s} deeplabv3 mimarisi ile anlamsal b{\"o}l{\"u}tleme",

abstract = "In this work, semantic segmentation has been dealt with convolutional neural networks (CNN) which is a widely used recent approach in the field of computer vision. In the experiments using Cityscapes dataset, the images are scaled by various rates and the CNN architecture named DeepLabv3 is trained with different hyperparameters using these images. After the training phase, the success rates of the trained models were compared. The most successful DeepLabv3 model has achieved a success rate of 78.83% on Cityscapes test set. Afterwards, an ensemble of two different DeepLabv3 models and the Extended DeepLabv3 model is tested. In test results, while the success rate remains nearly the same, an increase in classes such as road and sidewalk is observed.",

keywords = "Convolutional neural networks (CNN), Deep learning, Semantic segmentation",

author = "Yurtkulu, {Salih Can} and Sahin, {Yusuf Huseyin} and Gozde Unal",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 27th Signal Processing and Communications Applications Conference, SIU 2019 ; Conference date: 24-04-2019 Through 26-04-2019",

year = "2019",

month = apr,

doi = "10.1109/SIU.2019.8806244",

language = "T{\"u}rk{\c c}e",

series = "27th Signal Processing and Communications Applications Conference, SIU 2019",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "27th Signal Processing and Communications Applications Conference, SIU 2019",

address = "United States",

}

Yurtkulu, SC, Sahin, YH & Unal, G 2019, Genişletilmiş deeplabv3 mimarisi ile anlamsal bölütleme. in 27th Signal Processing and Communications Applications Conference, SIU 2019., 8806244, 27th Signal Processing and Communications Applications Conference, SIU 2019, Institute of Electrical and Electronics Engineers Inc., 27th Signal Processing and Communications Applications Conference, SIU 2019, Sivas, Turkey, 24/04/19. https://doi.org/10.1109/SIU.2019.8806244

Genişletilmiş deeplabv3 mimarisi ile anlamsal bölütleme. / Yurtkulu, Salih Can; Sahin, Yusuf Huseyin ; Unal, Gozde.
27th Signal Processing and Communications Applications Conference, SIU 2019. Institute of Electrical and Electronics Engineers Inc., 2019. 8806244 (27th Signal Processing and Communications Applications Conference, SIU 2019).

Araştırma sonucu: Kitap/Rapor/Konferans Bildirisinde Bölüm › Konferans katkısı › bilirkişi

TY - GEN

T1 - Genişletilmiş deeplabv3 mimarisi ile anlamsal bölütleme

AU - Yurtkulu, Salih Can

AU - Sahin, Yusuf Huseyin

AU - Unal, Gozde

PY - 2019/4

Y1 - 2019/4

N2 - In this work, semantic segmentation has been dealt with convolutional neural networks (CNN) which is a widely used recent approach in the field of computer vision. In the experiments using Cityscapes dataset, the images are scaled by various rates and the CNN architecture named DeepLabv3 is trained with different hyperparameters using these images. After the training phase, the success rates of the trained models were compared. The most successful DeepLabv3 model has achieved a success rate of 78.83% on Cityscapes test set. Afterwards, an ensemble of two different DeepLabv3 models and the Extended DeepLabv3 model is tested. In test results, while the success rate remains nearly the same, an increase in classes such as road and sidewalk is observed.

AB - In this work, semantic segmentation has been dealt with convolutional neural networks (CNN) which is a widely used recent approach in the field of computer vision. In the experiments using Cityscapes dataset, the images are scaled by various rates and the CNN architecture named DeepLabv3 is trained with different hyperparameters using these images. After the training phase, the success rates of the trained models were compared. The most successful DeepLabv3 model has achieved a success rate of 78.83% on Cityscapes test set. Afterwards, an ensemble of two different DeepLabv3 models and the Extended DeepLabv3 model is tested. In test results, while the success rate remains nearly the same, an increase in classes such as road and sidewalk is observed.

KW - Convolutional neural networks (CNN)

KW - Deep learning

KW - Semantic segmentation

UR - http://www.scopus.com/inward/record.url?scp=85071980720&partnerID=8YFLogxK

U2 - 10.1109/SIU.2019.8806244

DO - 10.1109/SIU.2019.8806244

M3 - Konferans katkısı

AN - SCOPUS:85071980720

T3 - 27th Signal Processing and Communications Applications Conference, SIU 2019

BT - 27th Signal Processing and Communications Applications Conference, SIU 2019

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 27th Signal Processing and Communications Applications Conference, SIU 2019

Y2 - 24 April 2019 through 26 April 2019

ER -

Genişletilmiş deeplabv3 mimarisi ile anlamsal bölütleme

Özet

Yayın serisi

???event.eventtypes.event.conference???

Bibliyografik not

Keywords

Belgeye Erişim

Diğer Dosyalar ve Linkler

Parmak izi

Alıntı Yap