Yuan Li
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
Yuan Li has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Hong Huang
Key Laboratory of Optoelectronic Technology and Systems of the Education Ministry of China, Chongqing University, Chongqing, China
4 shared publications - Qijuan Tan
Department of Radiology, Chongqing University Cancer Hospital & Chongqing Cancer Institute & Chongqing Cancer Hospital, Chongqing, China
2 shared publications - Yiming Tang
Key Laboratory of Optoelectronic Technology and Systems of the Education Ministry of China, Chongqing University, Chongqing, China
2 shared publications - Changyu Liang
Department of Radiology, Chongqing University Cancer Hospital & Chongqing Cancer Institute & Chongqing Cancer Hospital, Chongqing, China
1 shared publication - Chengli Song
University of Shanghai for Science and Technology
1 shared publication - Chunyu Pu
Key Laboratory of Optoelectronic Technology and Systems of the Education Ministry of China, Chongqing University, Chongqing, China
1 shared publication
Latest Publications
- 24 January 2026
- Published by Springer Nature in Applied Intelligence
- Vol. 56 (2) , 56
Collaborative deep learning framework based on adaptive feature fusion for malignancy prediction of lung nodules
- 24 January 2026
- Published by Springer Nature in Applied Intelligence
- Vol. 56 (2) , 56
Collaborative deep learning framework based on adaptive feature fusion for malignancy prediction of lung nodules
- 8 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 30 (10709908) , 543-547
Masked Spectral Bands Modeling With Shifted Windows: An Excellent Self-Supervised Learner for Classification of Medical Hyperspectral Images
Hyperspectral imaging has become a popular imaging technique in the medical field, and the development of algorithms for computer-aided diagnosis (CAD) is urgently required. Traditional deep learning techniques require a lot of annotated data, which is a burden on doctors. Self-supervised learning (SSL) is a solution for extracting feature representations from unlabeled data. However, traditional CNN-based SSL algorithms cannot explore relations between neighboring and long-range spectral bands, which limits classification performance. In this letter, the proposed solution is a novel SSL method using a transformer-based technique called masked spectral bands modeling with shifted windows (MSBMSW). This method predicts masked spectral bands as the pretext task and uses a self-attention mechanism with shifted windows to capture the divergence of neighboring spectral bands and enhance information exchange between long-range spectral bands. Experimental results demonstrate that MSBMSW achieves better classification results than many state-of-the-art methods and has potential clinical value for CAD of MHSIs.
- 8 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 30 (10709908) , 543-547
Masked Spectral Bands Modeling With Shifted Windows: An Excellent Self-Supervised Learner for Classification of Medical Hyperspectral Images
Hyperspectral imaging has become a popular imaging technique in the medical field, and the development of algorithms for computer-aided diagnosis (CAD) is urgently required. Traditional deep learning techniques require a lot of annotated data, which is a burden on doctors. Self-supervised learning (SSL) is a solution for extracting feature representations from unlabeled data. However, traditional CNN-based SSL algorithms cannot explore relations between neighboring and long-range spectral bands, which limits classification performance. In this letter, the proposed solution is a novel SSL method using a transformer-based technique called masked spectral bands modeling with shifted windows (MSBMSW). This method predicts masked spectral bands as the pretext task and uses a self-attention mechanism with shifted windows to capture the divergence of neighboring spectral bands and enhance information exchange between long-range spectral bands. Experimental results demonstrate that MSBMSW achieves better classification results than many state-of-the-art methods and has potential clinical value for CAD of MHSIs.
- 25 August 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (3) , 2381-2397
A Siamese network-based tracking framework for hyperspectral video
- 25 August 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (3) , 2381-2397
A Siamese network-based tracking framework for hyperspectral video
- Open Access
- 2 April 2021
- Published by MDPI AG in Remote Sensing
- Vol. 13 (7) , 1363
Dimensionality Reduction of Hyperspectral Image Based on Local Constrained Manifold Structure Collaborative Preserving Embedding
Graph learning is an effective dimensionality reduction (DR) manner to analyze the intrinsic properties of high dimensional data, it has been widely used in the fields of DR for hyperspectral image (HSI) data, but they ignore the collaborative relationship between sample pairs. In this paper, a novel supervised spectral DR method called local constrained manifold structure collaborative preserving embedding (LMSCPE) was proposed for HSI classification. At first, a novel local constrained collaborative representation (CR) model is designed based on the CR theory, which can obtain more effective collaborative coefficients to characterize the relationship between samples pairs. Then, an intraclass collaborative graph and an interclass collaborative graph are constructed to enhance the intraclass compactness and the interclass separability, and a local neighborhood graph is constructed to preserve the local neighborhood structure of HSI. Finally, an optimal objective function is designed to obtain a discriminant projection matrix, and the discriminative features of various land cover types can be obtained. LMSCPE can characterize the collaborative relationship between sample pairs and explore the intrinsic geometric structure in HSI. Experiments on three benchmark HSI data sets show that the proposed LMSCPE method is superior to the state-of-the-art DR methods for HSI classification.
- 2 April 2021
- Published by MDPI AG in Remote Sensing
- Vol. 13 (7) , 1363
Open AccessDimensionality Reduction of Hyperspectral Image Based on Local Constrained Manifold Structure Collaborative Preserving Embedding
Graph learning is an effective dimensionality reduction (DR) manner to analyze the intrinsic properties of high dimensional data, it has been widely used in the fields of DR for hyperspectral image (HSI) data, but they ignore the collaborative relationship between sample pairs. In this paper, a novel supervised spectral DR method called local constrained manifold structure collaborative preserving embedding (LMSCPE) was proposed for HSI classification. At first, a novel local constrained collaborative representation (CR) model is designed based on the CR theory, which can obtain more effective collaborative coefficients to characterize the relationship between samples pairs. Then, an intraclass collaborative graph and an interclass collaborative graph are constructed to enhance the intraclass compactness and the interclass separability, and a local neighborhood graph is constructed to preserve the local neighborhood structure of HSI. Finally, an optimal objective function is designed to obtain a discriminant projection matrix, and the discriminative features of various land cover types can be obtained. LMSCPE can characterize the collaborative relationship between sample pairs and explore the intrinsic geometric structure in HSI. Experiments on three benchmark HSI data sets show that the proposed LMSCPE method is superior to the state-of-the-art DR methods for HSI classification.
- Open Access
- 27 May 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 13 (19391404) , 2520-2531
Adaptive Residual Convolutional Neural Network for Hyperspectral Image Classification
In this paper, we designed an adaptive residual convolutional neural network (ARCNN) that takes raw hyperspectral image (HSI) cubes as input data for land cover classification. In this network, spectral and spatial feature extraction blocks are explored to learn discriminative features from abundant spectral information and spatial contexts in HSI, respectively. The proposed ARCNN is an end-to-end deep learning framework that alleviates the declining-accuracy phenomenon of deep learning models, and it also ranks the correlation and importance of each band in HSI. Furthermore, the residual blocks connect every other 3-D convolutional layer by using identity mapping, which facilitates the backpropagation of gradients. In order to address the common issue of imbalance between high dimensionality and limited availability of training samples for HSI classification, attention mechanism and feature fusion block are investigated to improve the performance of the ARCNN. Finally, some strategies, batch normalization and dropout are imposed on every convolutional layer to regularize the learning process. Therefore, the ARCNN method brings benefits to extract discriminative features, and it is easier to avoid over-fitting and achieves better performance. Experimental results on three public HSI data sets demonstrate the effectiveness of ARCNN over some state-of-the-art methods.
- 27 May 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 13 (19391404) , 2520-2531
Open AccessAdaptive Residual Convolutional Neural Network for Hyperspectral Image Classification
In this paper, we designed an adaptive residual convolutional neural network (ARCNN) that takes raw hyperspectral image (HSI) cubes as input data for land cover classification. In this network, spectral and spatial feature extraction blocks are explored to learn discriminative features from abundant spectral information and spatial contexts in HSI, respectively. The proposed ARCNN is an end-to-end deep learning framework that alleviates the declining-accuracy phenomenon of deep learning models, and it also ranks the correlation and importance of each band in HSI. Furthermore, the residual blocks connect every other 3-D convolutional layer by using identity mapping, which facilitates the backpropagation of gradients. In order to address the common issue of imbalance between high dimensionality and limited availability of training samples for HSI classification, attention mechanism and feature fusion block are investigated to improve the performance of the ARCNN. Finally, some strategies, batch normalization and dropout are imposed on every convolutional layer to regularize the learning process. Therefore, the ARCNN method brings benefits to extract discriminative features, and it is easier to avoid over-fitting and achieves better performance. Experimental results on three public HSI data sets demonstrate the effectiveness of ARCNN over some state-of-the-art methods.
- Open Access
- 1 January 2019
- Published by Taylor & Francis in International Journal of Hyperthermia
- Vol. 36 (1) , 1296-1305
Preliminary study of a control algorithm for radio-frequency-induced intestinal tissue fusion
Purpose: A control algorithm for radio-frequency-induced intestinal tissue fusion was developed to explore the effects of different control parameters on intestinal tissue fusion. Materials and methods: Radio-frequency-induced fusion was performed on ex vivo small intestine tissue. The effect on the fusion was observed by changing the control parameters (power, interval time, and terminal impedance) in the algorithm. The quality of fusion was evaluated using the burst pressure and thermal damage measurement. Histological evaluation was used to assess the fusion quality indirectly. Results: A maximum burst pressure of 8.460 ± 0.2674 KPa was acquired when the power was set to 100 W, the interval time was set to 2000 ms, and the terminal impedance was set to 50 Ω. Moreover, the thermal damage range increased with an increase in power but decreased with an increase in the interval time and terminal impedance. Furthermore, the thermal damage range and temperature were presumably related. Conclusions: For an ex vivo small intestine tissue, the appropriate control parameters could be set when the power was approximately 100 W, the interval time was approximately 2000 ms, and the terminal impedance was approximately 50 Ω. This study could provide a basis for the selection of control parameters for intestinal tissue fusion.
- 1 January 2019
- Published by Taylor & Francis in International Journal of Hyperthermia
- Vol. 36 (1) , 1296-1305
Open AccessPreliminary study of a control algorithm for radio-frequency-induced intestinal tissue fusion
Purpose: A control algorithm for radio-frequency-induced intestinal tissue fusion was developed to explore the effects of different control parameters on intestinal tissue fusion. Materials and methods: Radio-frequency-induced fusion was performed on ex vivo small intestine tissue. The effect on the fusion was observed by changing the control parameters (power, interval time, and terminal impedance) in the algorithm. The quality of fusion was evaluated using the burst pressure and thermal damage measurement. Histological evaluation was used to assess the fusion quality indirectly. Results: A maximum burst pressure of 8.460 ± 0.2674 KPa was acquired when the power was set to 100 W, the interval time was set to 2000 ms, and the terminal impedance was set to 50 Ω. Moreover, the thermal damage range increased with an increase in power but decreased with an increase in the interval time and terminal impedance. Furthermore, the thermal damage range and temperature were presumably related. Conclusions: For an ex vivo small intestine tissue, the appropriate control parameters could be set when the power was approximately 100 W, the interval time was approximately 2000 ms, and the terminal impedance was approximately 50 Ω. This study could provide a basis for the selection of control parameters for intestinal tissue fusion.