Algorithms for Synthetic Aperture Radar Imagery XXVI 2019
DOI: 10.1117/12.2523460
|View full text |Cite
|
Sign up to set email alerts
|

A SAR dataset for ATR Development: the Synthetic and Measured Paired Labeled Experiment (SAMPLE)

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

1
125
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
5
1
1

Relationship

1
6

Authors

Journals

citations
Cited by 96 publications
(126 citation statements)
references
References 19 publications
1
125
0
Order By: Relevance
“…We calculated the MSSIM index and the PSNR value by comparing the simulated SAR images and the MSTAR images to compare the similarity between the simulated and measured images. In the simulation, we used parameters identical to those of the MSTAR measurement parameters (Table VII) [42]. The same numbers of simulated and measured SAR images obtained from the same aspect angles and depression angles were used in each case when computing the MSSIM values corresponding to each model.…”
Section: Resultsmentioning
confidence: 99%
See 1 more Smart Citation
“…We calculated the MSSIM index and the PSNR value by comparing the simulated SAR images and the MSTAR images to compare the similarity between the simulated and measured images. In the simulation, we used parameters identical to those of the MSTAR measurement parameters (Table VII) [42]. The same numbers of simulated and measured SAR images obtained from the same aspect angles and depression angles were used in each case when computing the MSSIM values corresponding to each model.…”
Section: Resultsmentioning
confidence: 99%
“…This is presumably why it is occasionally adopted for the evaluation of simulated target SAR images [32] or for target recognition [36]. One study used other similarity measures such as mean absolute deviation, root mean square distance, and normalized cross correlation using the simulated and measured paired and labeled experiment dataset as a companion to the MSTAR dataset; however, the experiments were constrained and did not suggest the relevance to the recognition rate [34], [42].…”
Section: Related Workmentioning
confidence: 99%
“…Finally, we consider transfer learning with the Synthetic and Measured Paired and Labeled (SAMPLE) database of computer-simulated and real-world SAR images [13]. The publiclyavailable SAMPLE database consists of 1366 paired images of 10 different vehicles, each pair consisting of a real-world SAR image and a corresponding computer-simulated SAR image; see Figure 5 for an illustration.…”
Section: Transfer Learning With the Sample Datasetmentioning
confidence: 99%
“…For each technique, if test images cannot be represented in a feature space that has been determined from the training set, then classification accuracy is poor.In applications such as synthetic aperture radar (SAR) automatic target recognition (ATR), it is infeasible to collect the volume of data necessary to naively train high-accuracy classification networks. Additionally, due to varying operating conditions, the features measured in SAR imagery are different from those extracted from electro-optical (EO) imagery [13]. As such, off-the-shelf networks that have been pre-trained on the popular EObased ImageNet [3] or CIFAR-10 [8] datasets are insufficient for performing accurate ATR tasks in different imaging domains.…”
mentioning
confidence: 99%
“…For the next transfer task, we consider the synthetic-to-measured transfer task where our source domain consists of 10 classes of synthetic SAR images from the SAMPLE dataset 33 and our target domain consists of corresponding measured SAR images. The results are shown in Table 7.…”
Section: Experiments 4: Synthetic To Measured Sarmentioning
confidence: 99%