Efficient densely connected convolutional neural networks

•Proposed two efficient densely connected ConvNets, DesneDsc and Dense2Net.•DesneDsc and Dense2Net are more efficient and have higher accuracy than DenseNet.•Dense2Net are state-of-the-art on ImageNet in manual CNNs with

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	Pattern recognition 2021-01, Vol.109, p.107610, Article 107610
Hauptverfasser:	Li, Guoqing, Zhang, Meng, Li, Jiaojie, Lv, Feng, Tong, Guodong
Format:	Artikel
Sprache:	eng
Schlagworte:	Classification Computer Science Computer Science, Artificial Intelligence Convolutional neural networks Densely connected Engineering Engineering, Electrical & Electronic Parameter efficiency Science & Technology Technology
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	•Proposed two efficient densely connected ConvNets, DesneDsc and Dense2Net.•DesneDsc and Dense2Net are more efficient and have higher accuracy than DenseNet.•Dense2Net are state-of-the-art on ImageNet in manual CNNs with
ISSN:	0031-3203 1873-5142
DOI:	10.1016/j.patcog.2020.107610