TY - GEN
T1 - PDCapsNet
T2 - 13th International Conference on Information Technology in Medicine and Education, ITME 2023
AU - Liang, Zhihao
AU - Lu, Huijuan
AU - You, Cunqian
AU - Zhu, Wenjie
AU - Xie, Li
AU - Yao, Yudong
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - The COVID-19 epidemic started in 2019 and it is still important to develop intelligent and efficient diagnostic methods for the detection of COVID-19. In the task of COVID-19 X-ray image recognition, conventional deep neural networks encounter challenges in effectively discerning distinctive features within pneumonia images, resulting in limited generalization capability when confronted with novel samples. To address this issue, we introduce a novel model in this paper, denoted as PDCapsNet. The model effectively exploits Pyconv's multi-scale fusion technique and leverages the capsule network's capacity to encapsulate diverse attributes of specific entities within COVID-19 X-ray images, resulting in enhancements in both accuracy and generalization. Experimental results show that the proposed model outperforms other previous related works.
AB - The COVID-19 epidemic started in 2019 and it is still important to develop intelligent and efficient diagnostic methods for the detection of COVID-19. In the task of COVID-19 X-ray image recognition, conventional deep neural networks encounter challenges in effectively discerning distinctive features within pneumonia images, resulting in limited generalization capability when confronted with novel samples. To address this issue, we introduce a novel model in this paper, denoted as PDCapsNet. The model effectively exploits Pyconv's multi-scale fusion technique and leverages the capsule network's capacity to encapsulate diverse attributes of specific entities within COVID-19 X-ray images, resulting in enhancements in both accuracy and generalization. Experimental results show that the proposed model outperforms other previous related works.
KW - Chest X-ray images
KW - COVID-19
KW - Deep learning
KW - Image classification
UR - http://www.scopus.com/inward/record.url?scp=85192455277&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85192455277&partnerID=8YFLogxK
U2 - 10.1109/ITME60234.2023.00058
DO - 10.1109/ITME60234.2023.00058
M3 - Conference contribution
AN - SCOPUS:85192455277
T3 - Proceedings - 2023 13th International Conference on Information Technology in Medicine and Education, ITME 2023
SP - 245
EP - 249
BT - Proceedings - 2023 13th International Conference on Information Technology in Medicine and Education, ITME 2023
Y2 - 24 November 2023 through 26 November 2023
ER -