Yunlong Yu

Mr. Yunlong Yu

orcid

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 September 2019 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
YYYunlong Yu
Yunlong Yu
XLXianzhi Li
Xianzhi Li
FLFuxian Liu
Fuxian Liu
With the development of deep learning, supervised feature learning methods have achieved prominent performance in the field of aerial scene classification. However, supervised feature learning methods require a large amount of labeled training data. To address this limitation, in this article, a novel unsupervised deep feature learning method, namely, Attention generative adversarial networks (Attention GANs), is proposed for aerial scene classification. First, Attention GANs integrates the attention mechanism into GANs to enhance the representation power of the discriminator. Then, to obtain contextual information, a context-aggregation-based feature fusion architecture is designed in the discriminator. Furthermore, the generator and discriminator losses are improved on basis of the Relativistic GAN. At the same time, a content loss is formed by using the feature representations from the context-aggregation-based feature fusion architecture. In the experiments, our Attention GANs is evaluated via comprehensive experiments with four publicly available remote sensing scene data sets, i.e., the UC-Merced data set with 21 scene classes, the RSSCN7 data set with 7 scene classes, the AID data set with 30 scene classes, and the NWPU-RESISC45 data set with 45 scene classes. Experimental results demonstrate that our Attention GANs can obtain the best performance compared with the state-of-the-art methods.

23 July 2018 in Remote Sensing
Verified
Journal Article
Open access
PDF file
YYYunlong Yu
Yunlong Yu
FLFuxian Liu
Fuxian Liu
Aerial scene classification is an active and challenging problem in high-resolution remote sensing imagery understanding. Deep learning models, especially convolutional neural networks (CNNs), have achieved prominent performance in this field. The extraction of deep features from the layers of a CNN model is widely used in these CNN-based methods. Although the CNN-based approaches have obtained great success, there is still plenty of room to further increase the classification accuracy. As a matter of fact, the fusion with other features has great potential for leading to the better performance of aerial scene classification. Therefore, we propose two effective architectures based on the idea of feature-level fusion. The first architecture, i.e., texture coded two-stream deep architecture, uses the raw RGB network stream and the mapped local binary patterns (LBP) coded network stream to extract two different sets of features and fuses them using a novel deep feature fusion model. In the second architecture, i.e., saliency coded two-stream deep architecture, we employ the saliency coded network stream as the second stream and fuse it with the raw RGB network stream using the same feature fusion model. For sake of validation and comparison, our proposed architectures are evaluated via comprehensive experiments with three publicly available remote sensing scene datasets. The classification accuracies of saliency coded two-stream architecture with our feature fusion model achieve 97.79%, 98.90%, 94.09%, 95.99%, 85.02%, and 87.01% on the UC-Merced dataset (50% and 80% training samples), the Aerial Image Dataset (AID) (20% and 50% training samples), and the NWPU-RESISC45 dataset (10% and 20% training samples), respectively, overwhelming state-of-the-art methods.

5 January 2018 in IEEE Geoscience and Remote Sensing Letters
Verified
Journal Article
Closed access
YYYunlong Yu
Yunlong Yu
FLFuxian Liu
Fuxian Liu
One of the challenging problems in understanding high-resolution remote sensing images is aerial scene classification. A well-designed feature extractor and classifier can improve classification accuracy. In this letter, we construct three different convolutional neural networks with different sizes of receptive field, respectively. More importantly, we further propose a multilevel fusion method, which can make judgment by incorporating different levels' information. The aerial image and two patches extracted from the image are fed to these three different networks, and then, a probability fusion model is established for final classification. The effectiveness of the proposed method is tested on a more challenging data set-AID that has 10000 high-resolution remote sensing images with 30 categories. Experimental results show that our multilevel fusion model gets a significant classification accuracy improvement over all state-of-the-art references.

30 May 2019 in IEEE Access
Verified
Journal Article
Open access
YYYunlong Yu
Yunlong Yu
FLFuxian Liu
Fuxian Liu

29 January 2020 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
YYYunlong Yu
Yunlong Yu
XLXianzhi Li
Xianzhi Li
FLFuxian Liu
Fuxian Liu
Generative adversarial networks (GANs) have been widely used to single image superresolution (SR) (SISR), and these GAN-based methods have achieved significant performance on natural images due to their abilities to generate realistic textures. However, previous GAN-based methods are of poor performance when applied to remote sensing scene image SR. In order to further enhance the visual quality, in this article, we present a GAN-based SISR method by proposing a novel generator, which is capable of generating perceptually pleasing remote sensing scene images. First, we design the enhanced deep back-projection network (E-DBPN) generators based on the architecture of the original DBPN and mainly make two modifications. The first one is to add the proposed enhanced residual channel attention module (ERCAM) into the original DBPN, which can keep good properties of the original input features but also has the ability to emphasize more important features and suppressing less useful features. The other is to replace the concatenation operation with the proposed sequential feature fusion module (SFFM) for dealing with the feature maps generated by different up-projection units discriminatorily. As for the training process, the E-DBPN generator is first trained using the mean squared error (MSE) loss. Next, in order to improve the perceptual quality of the recovered images, we employ the content loss and the adversarial loss to train our initialized generator network. Experiments show that our method achieves state-of-the-art performance compared to other SISR methods.

8 February 2018 in Neural Processing Letters
Verified
Journal Article
Closed access
PDF file
YYYunlong Yu
Yunlong Yu
FLFuxian Liu
Fuxian Liu
SMSheng Mao
Sheng Mao