Lele Xu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Lele Xu has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Ye Li
Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences, Beijing, People’s Republic of China
3 shared publications Jinzhong Xu
Key Laboratory of Space Utilization, Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences, Beijing, China
2 shared publications- Baokun Shi
Beijing University of Technology, China
1 shared publication - Hongfei Wang
Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences, Beijing, China
1 shared publication - Jiaxi Sun
Institute of Automation NLPR Chinese Academy of Sciences
1 shared publication - Jinzhong Xu
Key Laboratory of Space Utilization, Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences, Beijing 100094, P.R. China
1 shared publication
Latest Publications
- Open Access
- 27 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 12-17
Instance Segmentation of Open Vocabulary Images by Fusing Fast Fourier Transform
- 27 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 12-17
Open AccessInstance Segmentation of Open Vocabulary Images by Fusing Fast Fourier Transform
- 29 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-14
BCTNet: Bi-Branch Cross-Fusion Transformer for Building Footprint Extraction
Building footprint extraction in remote sensing remains challenging due to the diverse appearances of buildings and confusing scenarios. Recently, researchers have revealed that both the globality and locality are vitally important in building footprint extraction tasks and proposed to incorporate the local context and global long-range dependency in the segmentation models. However, the inadequate integration of the globality and locality still leads to incomplete, fake or missing extraction results. To alleviate these problems, a novel segmentation method named Bi-branch Cross-fusion Transformer Network (BCTNet) is proposed in this study. Two parallel branches of the convolutional encoder branch (CB) and the transformer encoder branch (TB) are designed to extract multi-scale feature maps. A concatenation-then-cross-fusion transformer block (CCTB) is put forward to integrate the locality from the CB and globality from the TB in a cross-fusion way at each stage of the encoding process. Then, an adaptive gating module (AGM) is proposed to gate the feature maps from the CCTB to strengthen the important features while suppressing the irrelevant interference information. After that, the segmentation results can be obtained through a simple decoding process. Comprehensive experiments on two benchmark datasets demonstrate that the proposed BCTNet can achieve superior performance compared to the current state-of-the-art (SOTA) segmentation methods.
- 29 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-14
BCTNet: Bi-Branch Cross-Fusion Transformer for Building Footprint Extraction
Building footprint extraction in remote sensing remains challenging due to the diverse appearances of buildings and confusing scenarios. Recently, researchers have revealed that both the globality and locality are vitally important in building footprint extraction tasks and proposed to incorporate the local context and global long-range dependency in the segmentation models. However, the inadequate integration of the globality and locality still leads to incomplete, fake or missing extraction results. To alleviate these problems, a novel segmentation method named Bi-branch Cross-fusion Transformer Network (BCTNet) is proposed in this study. Two parallel branches of the convolutional encoder branch (CB) and the transformer encoder branch (TB) are designed to extract multi-scale feature maps. A concatenation-then-cross-fusion transformer block (CCTB) is put forward to integrate the locality from the CB and globality from the TB in a cross-fusion way at each stage of the encoding process. Then, an adaptive gating module (AGM) is proposed to gate the feature maps from the CCTB to strengthen the important features while suppressing the irrelevant interference information. After that, the segmentation results can be obtained through a simple decoding process. Comprehensive experiments on two benchmark datasets demonstrate that the proposed BCTNet can achieve superior performance compared to the current state-of-the-art (SOTA) segmentation methods.
- Open Access
- 3 November 2022
- Published by Association for Computing Machinery (ACM)
- p. 33-40
Distributed Learning based on Asynchronized Discriminator GAN for remote sensing image segmentation
Remote sensing images are usually distributed in different departments and contain private information, so they normally cannot be available publicly. However, it is a trend to jointly use remote sensing images from different departments, because it normally enables the model to capture more information and remote sensing image analysis based on deep learning generally requires lots of training data. To address the above problem, in this paper, we apply a distributed asynchronized discriminator GAN framework (DGAN) to jointly learn remote sensing images from different client nodes. The DGAN is composed of multiple distributed discriminators and a central generator, and only the synthetic remote sensing images generated by the DGAN are used to train a semantic segmentation model. Based on DGAN, we establish an experimental platform composed of multiple different hosts, which adopts socket and multi-process technology to realize asynchronous communication between hosts, and visualize the training and testing process. During DGAN training, instead of original remote sensing images or convolutional network model information, only synthetic images, losses and labeled images are exchanged between nodes. Therefore, the DGAN well protects the privacy and security of the original remote sensing images. We verify the performance of the DGAN on three remote sensing image datasets (City-OSM, WHU and Kaggle Ship). In the experiments, we take different distributions of remote sensing images in client nodes into consideration. The experiments show that the DGAN has a great capacity for distributed remote sensing image learning without sharing the original remote sensing images or the convolutional network model. Moreover, compared with a centralized GAN trained on all remote sensing images collected from all client nodes, the DGAN can achieve almost the same performance in semantic segmentation tasks for remote sensing images.
- 3 November 2022
- Published by Association for Computing Machinery (ACM)
- p. 33-40
Open AccessDistributed Learning based on Asynchronized Discriminator GAN for remote sensing image segmentation
Remote sensing images are usually distributed in different departments and contain private information, so they normally cannot be available publicly. However, it is a trend to jointly use remote sensing images from different departments, because it normally enables the model to capture more information and remote sensing image analysis based on deep learning generally requires lots of training data. To address the above problem, in this paper, we apply a distributed asynchronized discriminator GAN framework (DGAN) to jointly learn remote sensing images from different client nodes. The DGAN is composed of multiple distributed discriminators and a central generator, and only the synthetic remote sensing images generated by the DGAN are used to train a semantic segmentation model. Based on DGAN, we establish an experimental platform composed of multiple different hosts, which adopts socket and multi-process technology to realize asynchronous communication between hosts, and visualize the training and testing process. During DGAN training, instead of original remote sensing images or convolutional network model information, only synthetic images, losses and labeled images are exchanged between nodes. Therefore, the DGAN well protects the privacy and security of the original remote sensing images. We verify the performance of the DGAN on three remote sensing image datasets (City-OSM, WHU and Kaggle Ship). In the experiments, we take different distributions of remote sensing images in client nodes into consideration. The experiments show that the DGAN has a great capacity for distributed remote sensing image learning without sharing the original remote sensing images or the convolutional network model. Moreover, compared with a centralized GAN trained on all remote sensing images collected from all client nodes, the DGAN can achieve almost the same performance in semantic segmentation tasks for remote sensing images.
- 29 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-14
Gated Spatial Memory and Centroid-Aware Network for Building Instance Extraction
Automatic building extraction from high-resolution remote sensing images plays an important role in many application fields, such as the urban planning and photogrammetry. However, the complex background and large variety in building appearances in high-resolution remote sensing images make the building instance extraction challenging. In this study, we propose a novel two-stage instance segmentation network named gated spatial memory and centroid-aware network (GSMC) to handle these problems. Two new modules, including a gated spatial memory module (GSM) and a centroid-aware head (CH), are developed in our GSMC. The GSM is a top-down spatial structure and semantic information transmission module, where two gates including an input gate and a state gate are designed to strengthen the important features and replenish the lacking information. The CH is a new task head for regressing the geometric center of each instance, which can help to promote the accurate and complete recognition for irregularly shaped buildings. Experiments on the WHU Aerial data set, the WHU Satellite data set, and the Massachusetts Building data set demonstrate that the proposed GSMC can achieve consistently superior performances when compared with the recent state-of-the-art deep learning methods.
- 29 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-14
Gated Spatial Memory and Centroid-Aware Network for Building Instance Extraction
Automatic building extraction from high-resolution remote sensing images plays an important role in many application fields, such as the urban planning and photogrammetry. However, the complex background and large variety in building appearances in high-resolution remote sensing images make the building instance extraction challenging. In this study, we propose a novel two-stage instance segmentation network named gated spatial memory and centroid-aware network (GSMC) to handle these problems. Two new modules, including a gated spatial memory module (GSM) and a centroid-aware head (CH), are developed in our GSMC. The GSM is a top-down spatial structure and semantic information transmission module, where two gates including an input gate and a state gate are designed to strengthen the important features and replenish the lacking information. The CH is a new task head for regressing the geometric center of each instance, which can help to promote the accurate and complete recognition for irregularly shaped buildings. Experiments on the WHU Aerial data set, the WHU Satellite data set, and the Massachusetts Building data set demonstrate that the proposed GSMC can achieve consistently superior performances when compared with the recent state-of-the-art deep learning methods.