Lixiang Ru

Lixiang Ru

Auto-generated profile, based on publicly available information

Lixiang Ru

Ant Group, Hangzhou, China

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
4 September 2019 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
BDBo Du
Bo Du
LRLixiang Ru
Lixiang Ru
Chen WuChen Wu
Chen Wu
Liangpei ZhangLiangpei Zhang
Liangpei Zhang
Change detection has been a hotspot in the remote sensing technology for a long time. With the increasing availability of multi-temporal remote sensing images, numerous change detection algorithms have been proposed. Among these methods, image transformation methods with feature extraction and mapping could effectively highlight the changed information and thus has a better change detection performance. However, the changes of multi-temporal images are usually complex, and the existing methods are not effective enough. In recent years, the deep network has shown its brilliant performance in many fields, including feature extraction and projection. Therefore, in this paper, based on the deep network and slow feature analysis (SFA) theory, we proposed a new change detection algorithm for multi-temporal remotes sensing images called deep SFA (DSFA). In the DSFA model, two symmetric deep networks are utilized for projecting the input data of bi-temporal imagery. Then, the SFA module is deployed to suppress the unchanged components and highlight the changed components of the transformed features. The change vector analysis pre-detection is employed to find unchanged pixels with high confidence as training samples. Finally, the change intensity is calculated with chi-square distance and the changes are determined by threshold algorithms. The experiments are performed on two real-world data sets and a public hyperspectral data set. The visual comparison and the quantitative evaluation have shown that DSFA could outperform the other state-of-the-art algorithms, including other SFA-based and deep learning methods.

25 November 2020 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
LRLixiang Ru
Lixiang Ru
BDBo Du
Bo Du
Chen WuChen Wu
Chen Wu
Classifying multi-temporal scene land-use categories and detecting their semantic scene-level changes for remote sensing imagery covering urban regions could straightly reflect the land-use transitions. Existing methods for scene change detection rarely focus on the temporal correlation of bi-temporal features, and are mainly evaluated on small scale scene change detection datasets. In this work, we proposed a CorrFusion module that fuses the highly correlated components in bi-temporal feature embeddings. We first extract the deep representations of the bi-temporal inputs with deep convolutional networks. Then the extracted features will be projected into a lower-dimensional space to extract the most correlated components and compute the instance-level correlation. The cross-temporal fusion will be performed based on the computed correlation in CorrFusion module. The final scene classification results are obtained with softmax layers. In the objective function, we introduced a new formulation to calculate the temporal correlation more efficiently and stably. The detailed derivation of backpropagation gradients for the proposed module is also given. Besides, we presented a much larger scale scene change detection dataset with more semantic categories and conducted extensive experiments on this dataset. The experimental results demonstrated that our proposed CorrFusion module could remarkably improve the multi-temporal scene classification and scene change detection results.

9 March 2022 in International Journal of Computer Vision
Verified
Journal Article
Closed access
PDF file
LRLixiang Ru
Lixiang Ru
BDBo Du
Bo Du
YZYibing Zhan
Yibing Zhan
Chen WuChen Wu
Chen Wu

13 January 2022 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
JYJingwen Yuan
Jingwen Yuan
LRLixiang Ru
Lixiang Ru
SWShugen Wang
Shugen Wang
Chen WuChen Wu
Chen Wu
Over the past decade, many excellent data sharing efforts have enriched the remote sensing scene classification (SC) methods. These datasets have achieved great success in complex high-level semantic information interpretation. However, most existing datasets are collected from standard and ungeoreferenced image patches for algorithm training and evaluation. These datasets do not fit for practical applications and cannot be directly applied in further geographical study. Accordingly, we provide a large range high-resolution SC dataset with multiple time phases, called “ W u h an M ulti a pplication V HR S cene classification dataset (WH-MAVS).” It facilitates the study of SC and scene change detection (SCD) algorithms. Moreover, it can also be directly employed to perform a variety of real-life land use application tasks. To the best of our knowledge, this is the first free, publicly available, georeferenced, and annotated dataset to cover almost an entire megacity. The WH-MAVS was collected and annotated from Google Earth imagery with the same spatial resolution and uniform nonoverlapping patch size, covering the central area of Wuhan, China. The total number of scene samples is 47 137, which belong to 14 classes with 23 567 labeled patches for each time phase in 2014 and 2016, respectively. The geographic coordinates of all samples in both time phases exhibit one-to-one correspondence with 23 202 unchanged image patches of scene categories and 365 changed ones. The distribution of the number of samples in each class is highly imbalanced; moreover, there are large intraclass differences and indistinguishable interclass variances. These characteristics are closer to the real land use/land cover application tasks and introduce further challenges to the related algorithm research. In addition, we conducted benchmark experiments on SC and SCD based on the WH-MAVS dataset with widely used deep learning models. DenseNet169 was found to achieve the best performance. The overall accuracies are 91.07% and 92.09%, respectively, in the 2014 and 2016 validation sets of WH-MAVS. Furthermore, SCD obtained by DenseNet169 has a binary change detection accuracy of 89.56% and a multiple (from–to) change detection accuracy of 86.70%. Over and above the research value of the algorithm, it is also proven to have practical applications in fields such as urban planning, landscape pattern analysis, and urban dynamic monitoring and analysis.

25 February 2025 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
ZZZhenghui Zhao
Zhenghui Zhao
LRLixiang Ru
Lixiang Ru
Chen WuChen Wu
Chen Wu
DWDi Wang
Di Wang
Change detection (CD) based on deep learning typically requires costly pixel-level change labels. Recently, Weakly-Supervised Change Detection (WSCD) has emerged as a more label-efficient approach, using scene-level (i.e., image-level) labels to identify pixel-level changes in bi-temporal images. With only scene-level labels, existing WSCD methods are typically trained as scene-level change classification models. However, these methods often suffer from label-prediction inconsistency, with false changes frequently predicted in unchanged scenes. To address this issue, we propose TransWCD-SA, an end-to-end classifier-predictor framework. TransWCD-SA consists of a hierarchical transformer-based TransWCD classifier and a Scene-Adaptive (SA) predictor. This classifier-predictor framework is trained with two-stage joint constraints in an end-to-end learning manner. Specifically, the TransWCD classifier integrates hierarchical transformer blocks and multi-scale class activation maps, capturing pixel-level changes across various scales under weak supervision. The SA predictor dynamically introduces different pixel-level information for scenes labeled as changed and unchanged. Furthermore, a scene gated constraint is proposed as a penalty for label-prediction inconsistency, which is activated by the Dirac delta function and rectify features of mispredicted pixels in the embedding space. We validate the effectiveness of TransWCD-SA on three datasets: WHU-CD, LEVIR-CD, and DSIFN-CD, demonstrating significant improvement. The code is available at https://github.com/zhenghuizhao/TransWCD.

28 October 2024
Verified
Conference Paper
Open access
LRLixiang Ru
Lixiang Ru
XGXin Guo
Xin Guo
LYLei Yu
Lei Yu
YZYingying Zhang
Yingying Zhang
JLJiangwei Lao
Jiangwei Lao
MYMing Yang
Ming Yang
Long-tailed recognition (LTR) aims to learn balanced models from extremely unbalanced training data. Fine-tuning pretrained foundation models has recently emerged as a promising research direction for LTR. However, we observe that the fine-tuning process tends to degrade the intrinsic representation capability of pretrained models and lead to model bias towards certain classes, thereby hindering the overall recognition performance. To unleash the intrinsic representation capability of pretrained foundation models, in this work, we propose a new Parameter-Efficient Complementary Expert Learning (PECEL) for LTR. Specifically, PECEL consists of multiple experts, where individual experts are trained via Parameter-Efficient Fine-Tuning (PEFT) and encouraged to learn different expertise on complementary sub-categories via the proposed sample-aware logit adjustment loss. By aggregating the predictions of different experts, PECEL effectively achieves a balanced performance on long-tailed classes. Nevertheless, learning multiple experts generally introduces extra trainable parameters. To ensure parameter efficiency, we further propose a parameter sharing strategy which decomposes and shares the parameters in each expert. Extensive experiments on 4 LTR benchmarks show that the proposed PECEL can effectively learn multiple complementary experts without increasing the trainable parameters and achieve new state-of-the-art performance.