Unknown SAR Target Identification Method Based on Feature Extraction Network and KLD–RPA Joint Discrimination

Zeng, Zhiqiang; Sun, Jinping; Xu, Congan; Wang, Haiyang

doi:10.3390/rs13152901

Cited by 24 publications

(2 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If instead a code were used to distribute q, it would take on average H (p) + D (p||q) bits to describe the random variable. Relative entropy was first determined by Kullback and Leibler [29,30].…”

Section: The Methods Of Detecting Anomalies Based On Entropy In the Sensor Systemmentioning

confidence: 99%

An Adaptive Protection System for Sensor Networks Based on Analysis of Neighboring Nodes

Gamec

Basan

et al. 2021

Sensors

View full text Add to dashboard Cite

Creation and operation of sensor systems is a complex challenge not only for industrial and military purposes but also for consumer services (“smart city”, “smart home”) and other applications such as agriculture (“smart farm”, “smart greenhouse”). The use of such systems gives a positive economic effect and provides additional benefits from various points of view. At the same time, due to a large number of threats and challenges to cyber security, it is necessary to detect attacks on sensor systems in a timely manner. Here we present an anomaly detection method in which sensor nodes observe their neighbors and detect obvious deviations in their behavior. In this way, the community of neighboring nodes works collectively to protect one another. The nodes record only those parameters and attributes that are inherent in any node. Regardless of the node’s functionality, such parameters include the amount of traffic passing through the node, its Central Processing Unit (CPU) load, as well as the presence and number of packets dropped by the node. Our method’s main goal is to implement protection against the active influence of an internal attacker on the whole sensor network. We present the anomaly detection method, a dataset collection strategy, and experimental results that show how different types of attacks can be distinguished in the data produced by the nodes.

show abstract

Section: The Methods Of Detecting Anomalies Based On Entropy In the Sensor Systemmentioning

confidence: 99%

An Adaptive Protection System for Sensor Networks Based on Analysis of Neighboring Nodes

Gamec

Basan

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Ma et al [ 24 ] proposed an OSR method based on multitask learning, which was also developed from an adversarial network. Zeng et al [ 25 ] proposed the Fea-DA to finely identify the seen and unseen targets by calculating the relative position angle, which achieved the state-of-the-art (SOTA). Although these methods do not require prior knowledge, they require a large number of samples for training, which is not easy to achieve in reality either.…”

Section: Related Workmentioning

confidence: 99%

SAR Target Recognition with Limited Training Samples in Open Set Conditions

Zhou

Zhang

Liu

et al. 2023

Sensors

View full text Add to dashboard Cite

It is difficult to collect training samples for all types of synthetic aperture radar (SAR) targets. A realistic problem comes when unseen categories exist that are not included in training and benchmark data at the time of recognition, which is defined as open set recognition (OSR). Without the aid of side-information, generalized OSR methods used on ordinary optical images are usually not suitable for SAR images. In addition, OSR methods that require a large number of samples to participate in training are also not suitable for SAR images with the realistic situation of collection difficulty. In this regard, a task-oriented OSR method for SAR is proposed by distribution construction and relation measures to recognize targets of seen and unseen categories with limited training samples, and without any other simulation information. The method can judge category similarity to explain the unseen category. Distribution construction is realized by the graph convolutional network. The experimental results on the MSTAR dataset show that this method has a good recognition effect for the targets of both seen and unseen categories and excellent interpretation ability for unseen targets. Specifically, while recognition accuracy for seen targets remains above 95%, the recognition accuracy for unseen targets reaches 67% for the three-type classification problem, and 53% for the five-type classification problem.

show abstract

Evaluation of deep learning and transform domain feature extraction techniques for land cover classification: balancing through augmentation

Parikh

Patel

Patel³

2022

Environ Sci Pollut Res

View full text Add to dashboard Cite

Unknown SAR Target Identification Method Based on Feature Extraction Network and KLD–RPA Joint Discrimination

Cited by 24 publications

References 47 publications

An Adaptive Protection System for Sensor Networks Based on Analysis of Neighboring Nodes

An Adaptive Protection System for Sensor Networks Based on Analysis of Neighboring Nodes

SAR Target Recognition with Limited Training Samples in Open Set Conditions

Evaluation of deep learning and transform domain feature extraction techniques for land cover classification: balancing through augmentation

Contact Info

Product

Resources

About