Infrared aircraft few-shot classification method based on cross-correlation network

Fig. 5. (a) Training and validation accuracy curves of the baseline model and CCNet model on miniImageNet dataset；(b) Training and validation accuracy curves of the baseline model and CCNet model on miniInfra dataset

Download full size

View in Article

Fig. 6. Ablation experiment results on miniImageNet and miniInfra dataset

Download full size

View in Article

Fig. 7. The class activation mapping (CAM) feature visualization of CCNet

Download full size

View in Article

Table 1. Classification results on the miniImageNet dataset (average accuracy with 95% confidence interval)

View table

View in Article

Table 1. Classification results on the miniImageNet dataset (average accuracy with 95% confidence interval)

Method	Backbone	5-way 1-shot	5-way 5-shot
CCNet（ours）	ResNet12	66.20±0.43	81.82±0.31
MAML^［23］	ConvNet	48.70±0.84	63.11±0.92
RelationNet^［19］	ConvNet	50.44±0.82	65.32±0.70
CAN^［24］	ResNet12	63.85±0.48	79.44±0.34
AFHN^［25］	ResNet18	62.38±0.72	78.16±0.56
PSST^［26］	WRN-28-10	64.05±0.49	80.24±0.45
NCA^［27］	ResNet12	62.55±0.12	78.27±0.09
Mata-baseline^［28］	ResNet12	63.17±0.23	79.26±0.17
MIAN^［29］	ResNet12	64.27±0.35	81.24±0.26
TFH^［30］	ResNet18	64.49±0.84	79.94±0.60

Table 2. Classification results on the miniInfra dataset (average accuracy with 95% confidence interval)

View table

View in Article

Table 2. Classification results on the miniInfra dataset (average accuracy with 95% confidence interval)

Method	Pre-train	5-way 1-shot	5-way 5-shot	8-way 1-shot	8-way 5-shot
Improved Relation Network^［4］	No	84.37±1.31	93.66±0.76	77.56±1.46	90.58±0.64
Improved Relation Network^［4］	Yes	82.79±0.75	94.51±0.82	78.47±0.94	89.82±1.02
MLFC^［2］	No	—	—	78.58±0.97	91.12±0.37
MLFC^［2］	Yes	—	—	81.27±0.91	92.74±0.35
CCNet（ours）	No	85.58±0.97	95.09±0.46	81.95±0.62	93.26±0.38

Table 3. Comparison of accuracy and parameter quantities of different attention modules

View table

View in Article

Table 3. Comparison of accuracy and parameter quantities of different attention modules

Module	Self	Cross	miniImageNet	miniInfra	Add params
Baseline	√	×	64.86	81.27	0 k
SE^［31］	√	×	66.37	81.99	102.4 k
SCE^［32］	√	×	62.96	79.80	89.2 k
LSA^［33］	√	×	64.77	80.62	1 644.16 k
NLSA^［34］	√	×	65.67	82.34	822.1 k
CBAM^［35］	√	×	64.77	82.79	102.5 k
SCR^［9］	√	×	64.43	78.80	157.3 k
CCA^［9］	×	√	66.00	84.26	45.8 k
SAM	√	×	65.84	83.31	0 k
CA	×	√	65.69	84.30	9.41 k

Tools

Get Citation

Copy Citation Text

Zhen HUANG, Yong ZHANG, Jin-Fu GONG. Infrared aircraft few-shot classification method based on cross-correlation network[J]. Journal of Infrared and Millimeter Waves, 2025, 44(1): 96

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Infrared Optoelectronic System and Application Technology

Received: Mar. 29, 2024

Accepted: --

Published Online: Mar. 5, 2025

The Author Email: ZHANG Yong (zybxy@sina.com)

DOI:10.11972/j.issn.1001-9014.2025.01.013

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Classification results on the miniImageNet dataset (average accuracy with 95% confidence interval)

Table 1. Classification results on the miniImageNet dataset (average accuracy with 95% confidence interval)

Table 2. Classification results on the miniInfra dataset (average accuracy with 95% confidence interval)

Table 2. Classification results on the miniInfra dataset (average accuracy with 95% confidence interval)

Table 3. Comparison of accuracy and parameter quantities of different attention modules

Table 3. Comparison of accuracy and parameter quantities of different attention modules

微信扫一扫：分享