Multi-label infrared image classification algorithm based on weakly supervised learning

Table 1. Partial parameters of the experiment
View table
View in Article
Table 1. Partial parameters of the experiment
Name Parameter
Epoch 300
Batch_size 16
Optimizer Adam（lr=0.000 1）
lr_scheduler
StepLR
（Step_size=10 000，Gamma=0.5）

Table 2. Results of different algorithm

View table

View in Article

Table 2. Results of different algorithm

Model	Correct detection	False detection	Miss detection	F₁	Recall	Precision
ResNet50	77.84%	4.98%	17.18%	0.923 6	0.883 1	0.978 3
ResNet101	80.58%	7.56%	11.86%	0.936 0	0.961 6	0.920 1
Efficientnet	82.99%	6.36%	10.65%	0.942 5	0.920 7	0.978 9
Multi_path	82.65%	13.75%	3.61%	0.948 9	0.949 7	0.950 2
ResNeXt101	84.54%	13.57%	1.89%	0.948 0	0.947 9	0.949 8
C_train	90.38%	4.30%	5.33%	0.962 6	0.960 4	0.970 2
ML_IRI	91.24%	3.78%	4.98%	0.963 2	0.951 9	0.985 4

Table 3. Results of ablation experiment

View table

View in Article

Table 3. Results of ablation experiment

Model	Correct detection	False detection	Miss detection	F₁	Recall	Precision
ResNet50	77.84%	4.98%	17.18%	0.923 6	0.883 1	0.978 3
+ASL	81.10%	4.81%	14.09%	0.939 6	0.907 8	0.981 8
+CSRA	87.97%	2.75%	9.28%	0.951 5	0.926 3	0.989 9
ML_IRI	91.24%	3.78%	4.98%	0.963 2	0.951 9	0.985 4

Tools

Get Citation

Copy Citation Text

Chuankai MIAO, Shuli LOU, Ting LI, Huimin CAI. Multi-label infrared image classification algorithm based on weakly supervised learning[J]. Optics and Precision Engineering, 2022, 30(20): 2501

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites