Xinyu Yu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Xinyu Yu has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Honglin Wu
Changsha University of Science and Technology, Changsha, CHINA
3 shared publications - Min Zhang
School of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha, China
2 shared publications - Peng Huang
School of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha, China
2 shared publications - Zhaobin Zeng
School of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha, China
2 shared publications A.J. Pollack
Case Western Reserve University
1 shared publicationB. Drennan
Case Western Reserve University
1 shared publication
Latest Publications
- 26 November 2024
- Published by Springer Nature in The Visual Computer
- Vol. 41 (8) , 5391-5404
SSBFNet: a spectral-spatial fusion with BiFormer network for hyperspectral image classification
- 26 November 2024
- Published by Springer Nature in The Visual Computer
- Vol. 41 (8) , 5391-5404
SSBFNet: a spectral-spatial fusion with BiFormer network for hyperspectral image classification
- Open Access
- 28 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 17 (19391404) , 19986-19997
CCTNet: CNN and Cross-Shaped Transformer Hybrid Network for Remote Sensing Image Semantic Segmentation
Deep learning methods have achieved great success in the field of remote sensing image segmentation in recent years, but building a lightweight segmentation model with comprehensive local and global feature extraction capabilities remains a challenging task. In this paper, we propose a CNN and cross-shaped transformer hybrid network (CCTNet) for semantic segmentation of high-resolution remote sensing images. This model follows an encoder-decoder structure. It employs ResNet18 as an encoder to extract hierarchical feature information, and constructs a transformer decoder based on efficient cross-shaped self-attention (CSSA) to fully model local and global feature information and achieve lightweighting of the network. Moreover, the transformer block introduces a mixed-scale convolutional feedforward network (MCFFN) to further enhance multi-scale information extraction. Furthermore, a simplified and efficient feature aggregation module (FAM) is leveraged to gradually aggregate local and global information at different stages. Extensive comparison experiments on the ISPRS Vaihingen and Potsdam datasets reveal that our method obtains superior performance compared with state-of-the-art lightweight methods.
- 28 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 17 (19391404) , 19986-19997
Open AccessCCTNet: CNN and Cross-Shaped Transformer Hybrid Network for Remote Sensing Image Semantic Segmentation
Deep learning methods have achieved great success in the field of remote sensing image segmentation in recent years, but building a lightweight segmentation model with comprehensive local and global feature extraction capabilities remains a challenging task. In this paper, we propose a CNN and cross-shaped transformer hybrid network (CCTNet) for semantic segmentation of high-resolution remote sensing images. This model follows an encoder-decoder structure. It employs ResNet18 as an encoder to extract hierarchical feature information, and constructs a transformer decoder based on efficient cross-shaped self-attention (CSSA) to fully model local and global feature information and achieve lightweighting of the network. Moreover, the transformer block introduces a mixed-scale convolutional feedforward network (MCFFN) to further enhance multi-scale information extraction. Furthermore, a simplified and efficient feature aggregation module (FAM) is leveraged to gradually aggregate local and global information at different stages. Extensive comparison experiments on the ISPRS Vaihingen and Potsdam datasets reveal that our method obtains superior performance compared with state-of-the-art lightweight methods.
- Top Cited Papers
- 12 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-12
CMTFNet: CNN and Multiscale Transformer Fusion Network for Remote-Sensing Image Semantic Segmentation
Convolutional neural networks (CNNs) are powerful in extracting local information but lack the ability to model long-range dependencies. In contrast, transformer relies on multihead self-attention mechanisms to effectively extract the global contextual information and thus model long-range dependencies. In this paper, we propose a novel encoder-decoder structured semantic segmentation network, named as CNN and multiscale transformer fusion network (CMTFNet), to extract and fuse local information and multiscale global contextual information of high-resolution remote sensing images. Specifically, to further process the output features from the CNN encoder, we build a transformer decoder based on the multiscale multihead self-attention (M2SA) module for extracting rich multiscale global contextual information and channel information. Additionally, the transformer block introduces an efficient feed-forward network (E-FFN) to enhance the information interaction between different channels of the feature. Finally, the multiscale attention fusion (MAF) module fully fuses the feature information from different levels. We have conducted extensive comparison experiments and ablation experiments on the International Society for Photogrammetry and Remote Sensing (ISPRS) Vaihingen and Potsdam datasets. The extensive experimental results demonstrate that our proposed CMTFNet can obtain superior performance compared to the currently popular methods. The codes will be available at https://github.com/DrWuHonglin/CMTFNet.
- 12 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-12
Top Cited PapersCMTFNet: CNN and Multiscale Transformer Fusion Network for Remote-Sensing Image Semantic Segmentation
Convolutional neural networks (CNNs) are powerful in extracting local information but lack the ability to model long-range dependencies. In contrast, transformer relies on multihead self-attention mechanisms to effectively extract the global contextual information and thus model long-range dependencies. In this paper, we propose a novel encoder-decoder structured semantic segmentation network, named as CNN and multiscale transformer fusion network (CMTFNet), to extract and fuse local information and multiscale global contextual information of high-resolution remote sensing images. Specifically, to further process the output features from the CNN encoder, we build a transformer decoder based on the multiscale multihead self-attention (M2SA) module for extracting rich multiscale global contextual information and channel information. Additionally, the transformer block introduces an efficient feed-forward network (E-FFN) to enhance the information interaction between different channels of the feature. Finally, the multiscale attention fusion (MAF) module fully fuses the feature information from different levels. We have conducted extensive comparison experiments and ablation experiments on the International Society for Photogrammetry and Remote Sensing (ISPRS) Vaihingen and Potsdam datasets. The extensive experimental results demonstrate that our proposed CMTFNet can obtain superior performance compared to the currently popular methods. The codes will be available at https://github.com/DrWuHonglin/CMTFNet.
- Open Access
- 1 July 2022
- Published by KSI Research Inc.
- Vol. 2022, 628-633
Postoperative MPA-AUC0-12h Prediction for Kidney Transplant Recipients based on Few-shot Learning
- 1 July 2022
- Published by KSI Research Inc.
- Vol. 2022, 628-633
Open AccessPostoperative MPA-AUC0-12h Prediction for Kidney Transplant Recipients based on Few-shot Learning
- 27 March 2022
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12169, 66-360
Design and implementation of CCD imaging system for directional polarimetric camera
A CCD imaging system in the Directional Polarimetric Camera is developed, which is comprised of timing driving unit, pro-processing and analog-front-end unit, FPGA main control unit, internal communication and remote measurement unit, image transmission interface unit, and so on. A CCD driving timing method with twice frame transfer and once horizontal readout is proposed to effectively eliminate the residual charge of last frame image and improve the signal-to-noise ratio. The performance of this CCD imaging system is verified. The experimental results show that the 14bit image data can output steadily with a frame frequency of 2.02 frames per second. The imaging signal-to-noise ratio can reach 54.93dB when the CCD is lower than saturated. The dynamic range of the CCD detectable signals is 68.98 dB and the nonlinearity error under different wavelengths is less than 1%. The instrumentation can satisfy the operational requirements such as stable output, excellent linear performance, high signal-to-noise ratio and large dynamic range.
- 27 March 2022
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12169, 66-360
Design and implementation of CCD imaging system for directional polarimetric camera
A CCD imaging system in the Directional Polarimetric Camera is developed, which is comprised of timing driving unit, pro-processing and analog-front-end unit, FPGA main control unit, internal communication and remote measurement unit, image transmission interface unit, and so on. A CCD driving timing method with twice frame transfer and once horizontal readout is proposed to effectively eliminate the residual charge of last frame image and improve the signal-to-noise ratio. The performance of this CCD imaging system is verified. The experimental results show that the 14bit image data can output steadily with a frame frequency of 2.02 frames per second. The imaging signal-to-noise ratio can reach 54.93dB when the CCD is lower than saturated. The dynamic range of the CCD detectable signals is 68.98 dB and the nonlinearity error under different wavelengths is less than 1%. The instrumentation can satisfy the operational requirements such as stable output, excellent linear performance, high signal-to-noise ratio and large dynamic range.
- 13 November 2002
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 845-851 vol.2
A miniature hybrid robot propelled by legs
Describes the development of an autonomous hybrid micro-robot that uses legs for propulsion and support of the rear half of the body and uses a pair of wheels for support of the front half. McKibben artificial muscles actuate the legs and the compressed air that activates the actuators is generated by an on-board power plant comprising a pair of lithium batteries powering a gear motor driven air compressor. The control is also onboard in the form of a PIC that controls the actuators through four three-way valves that each consists of a pair of MEMS devices.
- 13 November 2002
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 845-851 vol.2
A miniature hybrid robot propelled by legs
Describes the development of an autonomous hybrid micro-robot that uses legs for propulsion and support of the rear half of the body and uses a pair of wheels for support of the front half. McKibben artificial muscles actuate the legs and the compressed air that activates the actuators is generated by an on-board power plant comprising a pair of lithium batteries powering a gear motor driven air compressor. The control is also onboard in the form of a PIC that controls the actuators through four three-way valves that each consists of a pair of MEMS devices.