SPNet: Siamese-Prototype Network for Few-Shot Remote Sensing Image Scene Classification
The DOI matches its Crossref record; 47 of the 47 references it cites were found in a registry or an index.
Abstract
Few-shot image classification has attracted extensive attention, which aims to recognize unseen classes given only a few labeled samples. Due to the large intraclass variances and interclass similarity of remote sensing scenes, the task under such circumstance is much more challenging than general few-shot image classification. Most existing prototype-based few-shot algorithms usually calculate prototypes directly from support samples and ignore the validity of prototypes, which results in a decline in the accuracy of subsequent inferences based on prototypes. To tackle this problem, we propose a Siamese-prototype network (SPNet) with prototype self-calibration (SC) and intercalibration (IC). First, to acquire more accurate prototypes, we utilize the supervision information from support labels to calibrate the prototypes generated from support features. This process is called SC. Second, we propose to consider the confidence scores of the query samples as another type of prototypes, which are then used to predict the support samples in the same way. Thus, the information interaction between support and query samples is implicitly a further calibration for prototypes (so-called IC). Our model is optimized with three losses, of which two additional losses help the model to learn more representative prototypes and make more accurate predictions. With no additional parameters to be learned, our model is very lightweight and convenient to employ. The experiments on three public remote sensing image datasets demonstrate competitive performance compared with other advanced few-shot image classification approaches. The source code is available athttps://github.com/zoraup/SPNet.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 2×
“We perform our method mainly on ResNet-18 [57] back-bone, and no other mechanisms or layers added.”
BackgroundMethodModel-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
Finn, C., Abbeel, P., & Levine, S. (2017) · arXiv · doi:10.48550/arxiv.1703.03400 · Cited by 8,512 · cited 3×
“Optimization-based methods consider that the common gradient descent methods are difficult to fit the few-shot scenario, so the task is carried out by adjusting the optimization method [27]–[30].”
MethodLearning to Compare: Relation Network for Few-Shot Learning
Sung, F., Yang, Y., Zhang, L., Xiang, T., Torr, P. H. S., & Hospedales, T. M. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00131 · Cited by 5,816
Bag-of-visual-words and spatial extensions for land-use classification
Yang, Y., & Newsam, S. (2010) · Proceedings of the 18th SIGSPATIAL International Conference on Advances in Geographic Information Systems · doi:10.1145/1869790.1869829 · Cited by 2,969 · cited 1×
“We carried out experiments on three of the most commonly used public datasets: NWPU-RESISC45 dataset [55], WHU-RS19 dataset [59], and UC Merced dataset [60].”
MethodRemote Sensing Image Scene Classification: Benchmark and State of the Art
Cheng, G., Han, J., & Lu, X. (2017) · Proceedings of the IEEE · doi:10.1109/jproc.2017.2675998 · Cited by 2,926 · cited 2×
“[16] devised a new remote sensing image scene classification framework named Hydra, an ensemble of CNNs which achieves excellent results on the NWPURESISC45 [55] dataset.”
MethodSelf-citationPANet: Few-Shot Image Semantic Segmentation With Prototype Alignment
Wang, K., Liew, J. H., Zou, Y., Zhou, D., & Feng, J. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00929 · Cited by 1,500 · cited 3×
“Referring to the common settings in some literature works [19], [37], we set β to 10, 20, and 50 and found that 50 performs better.”
BackgroundCheng, G., Yang, C., Yao, X., Guo, L., & Han, J. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2783902 · Cited by 1,226 · cited 1×
“Cheng et al. [9] designed a novel objec-tive function for learning discriminative CNNs (D-CNNs).”
BackgroundSelf-citationCheng, G., Xie, X., Han, J., Guo, L., & Xia, G.-S. (2020) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2020.3005403 · Cited by 995 · cited 1×
“One can refer to the recent survey paper [56] for more information about remote sensing image scene classification.”
BackgroundSelf-citationDeep learning based feature selection for remote sensing scene classification
Zou, Q., Ni, L., Zhang, T., & Wang, Q. (2015) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2015.2475299 · Cited by 956
Towards better exploiting convolutional neural networks for remote sensing scene classification
Nogueira, K., Penatti, O. A. B., & dos Santos, J. A. (2017) · Pattern Recognition · doi:10.1016/j.patcog.2016.07.001 · Cited by 902
Meta-SGD: Learning to Learn Quickly for Few-Shot Learning
Li, Z., Zhou, F., Chen, F., & Li, H. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1707.09835 · Cited by 763 · cited 3×
“…to other advanced methods to a great extent in most cases, including MatchingNet† [24], ProtoNet† [25], RelationNet† [26], MAML [27], Meta-SGD [30], LLSR [29], and DLA-MatchNet [36], where we use the same backbone as ours to reimplement MatchingNet†, ProtoNet†, and RelationNet† for fair…”
MethodDo deep features generalize from everyday objects to remote sensing and aerial scenes domains?
Penatti, O. A. B., Nogueira, K., & dos Santos, J. A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2015.7301382 · Cited by 705 · cited 1×
“In 2015, Penatti et al. [13] first proposed to use convolutional neural networks (CNNs) to deal with remote sensing image scene classification.”
BackgroundRevisiting Local Descriptor Based Image-To-Class Measure for Few-Shot Learning
Li, W., Wang, L., Xu, J., Huo, J., Gao, Y., & Luo, J. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00743 · Cited by 654
Deep Feature Fusion for VHR Remote Sensing Scene Classification
Chaib, S., Liu, H., Gu, Y., & Yao, H. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2700322 · Cited by 461 · cited 2×
“Chaib et al. [14] enhanced scene classification accuracy by fusing features extracted from CNNs.”
BackgroundZhao, B., Zhong, Y., Xia, G.-S., & Zhang, L. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2015.2496185 · Cited by 361 · cited 1×
“R EMOTE sensing image scene classification has been developing these years rapidly as it has a wide range of applications in the real world, such as natural hazards detection [1], [2], land-use classification [3]–[5], and urban planning [6], [7].”
BackgroundSemantic Annotation of High-Resolution Satellite Images via Weakly Supervised Learning
Yao, X., Han, J., Cheng, G., Qian, X., & Guo, L. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2016.2523563 · Cited by 346
Self-citationRemote Sensing Image Scene Classification Using Bag of Convolutional Features
Cheng, G., Li, Z., Yao, X., Guo, L., & Wei, Z. (2017) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2017.2731997 · Cited by 334
Self-citationMei, S., Ji, J., Geng, Y., Zhang, Z., Li, X., & Du, Q. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2019.2908756 · Cited by 315 · cited 1×
“Therefore, we then tried many combinations of λ 1 and λ 2 in the interval [1, 10].”
MethodMei, S., Ji, J., Hou, J., Li, X., & Du, Q. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2693346 · Cited by 311 · cited 1×
“” In recent years, deep learning has been a powerful means to accomplish many tasks in image and video processing [8]–[12].”
BackgroundSheng, G., Yang, W., Xu, T., & Sun, H. (2012) · International Journal of Remote Sensing · doi:10.1080/01431161.2011.608740 · Cited by 309 · cited 1×
“We carried out experiments on three of the most commonly used public datasets: NWPU-RESISC45 dataset [55], WHU-RS19 dataset [59], and UC Merced dataset [60].”
MethodTayyebi, A., Pijanowski, B. C., & Tayyebi, A. H. (2010) · Landscape and Urban Planning · doi:10.1016/j.landurbplan.2010.10.007 · Cited by 302 · cited 1×
“R EMOTE sensing image scene classification has been developing these years rapidly as it has a wide range of applications in the real world, such as natural hazards detection [1], [2], land-use classification [3]–[5], and urban planning [6], [7].”
BackgroundRemote Sensing Scene Classification by Unsupervised Representation Learning
Lu, X., Zheng, X., & Yuan, Y. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2702596 · Cited by 288
Remote Sensing Scene Classification Using Multilayer Stacked Covariance Pooling
He, N., Fang, L., Li, S., Plaza, A., & Plaza, J. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2845668 · Cited by 284 · cited 1×
“He et al. [15] proposed a novel multilayer stacked covariance pooling (MSCP) algo-rithm for remote sensing image scene classification, which can combine multilayer feature maps extracted from pretrained CNN automatically.”
BackgroundCheng, G., Guo, L., Zhao, T., Han, J., Li, H., & Fang, J. (2013) · International Journal of Remote Sensing · doi:10.1080/01431161.2012.705443 · Cited by 281 · cited 2×
“Therefore, we then tried many combinations of λ 1 and λ 2 in the interval [1, 10].”
BackgroundMethodSelf-citationPrototype Rectification for Few-Shot Learning
Liu, J., Song, L., & Qin, Y. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58452-8_43 · Cited by 269
Prototype-CNN for Few-Shot Object Detection in Remote Sensing Images
Cheng, G., Yan, B., Shi, P., Li, K., Yao, X., Guo, L., & Han, J. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3078507 · Cited by 193 · cited 1×
“Asthename implies, few-shot learning aims to learn new concepts from a few annotated examples [19], [20].”
BackgroundSelf-citationDLA-MatchNet for Few-Shot Remote Sensing Image Scene Classification
Li, L., Han, J., Yao, X., Cheng, G., & Guo, L. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3033336 · Cited by 192 · cited 10×
“Following closely behind, Li et al. [36] adopted the attention technique to discover discriminative regions in the remote sensing images automatically.”
BackgroundMethodInfluentialSelf-citationDeep Relation Network for Hyperspectral Image Few-Shot Classification
Gao, K., Liu, B., Yu, X., Qin, J., Zhang, P., & Tan, X. (2020) · Remote Sensing · doi:10.3390/rs12060923 · Cited by 183 · cited 1×
“For example, Gao et al. [43] proposed implementing the hyperspectral image classification in the case of few-shot learning.”
BackgroundScene Classification Based on Multiscale Convolutional Neural Network
Liu, Y., Zhong, Y., & Qin, Q. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2848473 · Cited by 181 · cited 1×
“[17] proposed the multiscale CNN (MCNN) to tackle the object scale variation problem in remote sensing scenes.”
BackgroundHydra: An Ensemble of Convolutional Neural Networks for Geospatial Land Classification
Minetto, R., Segundo, M. P., & Sarkar, S. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2019.2906883 · Cited by 162 · cited 1×
“Minetto et al. [16] devised a new remote sensing image scene classification framework named Hydra, an ensemble of CNNs which achieves excellent results on the NWPU-RESISC45 [55] dataset.”
MethodPrototype Completion with Primitive Knowledge for Few-Shot Learning
Zhang, B., Li, X., Ye, Y., Huang, Z., & Zhang, L. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00375 · Cited by 159 · cited 1×
“Among the classical few-shot classification methods, the prototypical network proposed by Snell et al. [25] is simple in principle but effective in practice, which has derived a series of studies based on it [34], [39]–[41].”
MethodMa, A., Wan, Y., Zhong, Y., Wang, J., & Zhang, L. (2021) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2020.11.025 · Cited by 158 · cited 1×
“They verified that the high-level features extracted by CNNs were more effective than the traditional low-level features, where-after a great deal of CNN-based scene classification methods appeared [14]–[18].”
BackgroundRS-MetaNet: Deep Metametric Learning for Few-Shot Remote Sensing Scene Classification
Li, H., Cui, Z., Zhu, Z., Chen, L., Zhu, J., Huang, H., & Tao, C. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3027387 · Cited by 144 · cited 2×
“Recently, Li et al. [33] introduced meta-learning to remote sensing image scene classification.”
BackgroundMethodCheng, G., Sun, X., Li, K., Guo, L., & Han, J. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3081421 · Cited by 130 · cited 1×
“As an important part of remote sensing image understanding, remote sensing image scene classification has a steady development [45]–[54].”
BackgroundSelf-citationLearning Deep Cross-Modal Embedding Networks for Zero-Shot Remote Sensing Image Scene Classification
Li, Y., Zhu, Z., Yu, J.-G., & Zhang, Y. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3047447 · Cited by 129 · cited 1×
“Lately, zero-shot classification was exploited in remote sensing images [44].”
BackgroundTransfer Learning With Fully Pretrained Deep Convolution Networks for Land-Use Classification
Zhao, B., Huang, B., & Zhong, Y. (2017) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2017.2691013 · Cited by 121
Multisource Compensation Network for Remote Sensing Cross-Domain Scene Classification
Lu, X., Gong, T., & Zheng, X. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2019.2951779 · Cited by 119
Zhong, Y., Su, Y., Wu, S., Zheng, Z., Zhao, J., Ma, A., Zhu, Q., Ye, R., Li, X., Pellikka, P., et al. (2020) · Remote Sensing of Environment · doi:10.1016/j.rse.2020.111838 · Cited by 116 · cited 1×
“R EMOTE sensing image scene classification has been developing these years rapidly as it has a wide range of applications in the real world, such as natural hazards detection [1], [2], land-use classification [3]–[5], and urban planning [6], [7].”
BackgroundFew-Shot Classification of Aerial Scene Images via Meta-Learning
Zhang, P., Bai, Y., Wang, D., Bai, B., & Li, Y. (2020) · Remote Sensing · doi:10.3390/rs13010108 · Cited by 109
Lifelong Learning for Scene Recognition in Remote Sensing Images
Zhai, M., Liu, H., & Sun, F. (2019) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2019.2897652 · Cited by 82 · cited 7×
“LLSR [29] is also an outstanding work on few-shot remote sensing image scene classification.”
BackgroundMethodInfluentialScene Classification Using Hierarchical Wasserstein CNN
Liu, Y., Suen, C. Y., Liu, Y., & Ding, L. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2873966 · Cited by 78 · cited 1×
“Liu et al. [11] introduced a hierarchical Wasseratein loss function to improve CNNs’ discriminatory ability.”
BackgroundTask-wise attention guided part complementary learning for few-shot image classification
Cheng, G., Li, R., Lang, C., & Han, J. (2021) · Science China Information Sciences · doi:10.1007/s11432-020-3156-7 · Cited by 65 · cited 1×
“Among these methods, metric-based methods are mostly studied in the past few years in both natural image classification and remote sensing image scene classification [24]–[26], [31]–[38].”
MethodSelf-citationMeta-Learning based prototype-relation network for few-shot classification
Liu, X., Zhou, F., Liu, J., & Jiang, L. (2019) · Neurocomputing · doi:10.1016/j.neucom.2019.12.034 · Cited by 65
Chen, J., Wang, C., Ma, Z., Chen, J., He, D., & Ackland, S. (2018) · Remote Sensing · doi:10.3390/rs10020290 · Cited by 65 · cited 1×
“As an important part of remote sensing image understanding, remote sensing image scene classification has a steady development [45]–[54].”
BackgroundSelf-citationScene Classification Based on the Sparse Homogeneous–Heterogeneous Topic Feature Model
Zhu, Q., Zhong, Y., Wu, S., Zhang, L., & Li, D. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2781712 · Cited by 62
Huang, X., Li, S., Li, J., Jia, X., Li, J., Zhu, X. X., & Benediktsson, J. A. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3037211 · Cited by 31 · cited 1×
“R EMOTE sensing image scene classification has been developing these years rapidly as it has a wide range of applications in the real world, such as natural hazards detection [1], [2], land-use classification [3]–[5], and urban planning [6], [7].”
BackgroundEmbedded adaptive cross-modulation neural network for few-shot learning
Wang, P., Cheng, J. S., Hao, F., Wang, L., & Feng, W. (2019) · Neural Computing and Applications · doi:10.1007/s00521-019-04605-y · Cited by 26 · cited 1×
“Among the classical few-shot classification methods, the prototypical network proposed by Snell et al. [25] is simple in principle but effective in practice, which has derived a series of studies based on it [34], [39]–[41].”
Method
Metadata
- Authors
- Cheng, G., Cai, L., Lang, C., Yao, X., Chen, J., Guo, L., & Han, J.
- Venue
- IEEE Transactions on Geoscience and Remote Sensing
- Published
- 2022
- Volume
- 60
- Pages
- 1-11
- ISSN
- 0196-2892 (ISSN-L)
- 1558-0644
- OpenAlex
- W3188824417
- Cited by
- 165
- References
- 47
- FWCI
- 10.76
- Type
- journal article
- Language
- en
- Open access
- closed
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.