Channel Attention Multi-Branch Network for Fine-Grained Image Recognition

Table 1. Ablation experimental results of our method on different datasets unit: %
View table
Table 1. Ablation experimental results of our method on different datasets unit: %
Method Stanford Cars Food-101
Without ECA sub-network 94.8 88.5
Without DO-Conv 95.2 89.8
Without AAPM 92.5 85.4
Ours 95.4 90.6

Table 2. Experimental results of different methods on the Stanford Cars dataset unit: %
View table
Table 2. Experimental results of different methods on the Stanford Cars dataset unit: %
Method Added training Accuracy
Bilinear-CNN(2015) × 91.3
RA-CNN(2017) × 92.5
HS-Net(2017) √ 93.8
AAA Model (2019) √ 95.4
MMAL(2020) × 95.0
Ours × 95.4

Table 3. Experimental results of different methods on the Food-101 dataset unit: %
View table
Table 3. Experimental results of different methods on the Food-101 dataset unit: %
Method Added training Accuracy
Bilinear-CNN(2015) × 84.7
WISeR(2018) × 90.3
FPCNN(2018) × 87.9
FCA(2019) √ 86.3
Ours × 90.6

Tools

Get Citation

Copy Citation Text

Binzhou Wang, Zhiyong Xiao. Channel Attention Multi-Branch Network for Fine-Grained Image Recognition[J]. Laser & Optoelectronics Progress, 2021, 58(22): 2210008

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Image Processing

Received: Jan. 4, 2021

Accepted: Jan. 27, 2021

Published Online: Nov. 5, 2021

The Author Email: Zhiyong Xiao (zhiyong.xiao@jiangnan.edu.cn)

DOI:10.3788/LOP202158.2210008

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Ablation experimental results of our method on different datasets unit: %

Table 1. Ablation experimental results of our method on different datasets unit: %

Table 2. Experimental results of different methods on the Stanford Cars dataset unit: %

Table 2. Experimental results of different methods on the Stanford Cars dataset unit: %

Table 3. Experimental results of different methods on the Food-101 dataset unit: %

Table 3. Experimental results of different methods on the Food-101 dataset unit: %