Wencheng 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
Research Subjects
Wencheng Li has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
Cong Peng
Nanchang Univ, Sch Informat Engn, Nanchang 330031, Jiangxi, Peoples R China
1 shared publication- Gang Chen
Sun Yat-sen University
1 shared publication - Qiurong Yan
Nanchang University
1 shared publication Sheng-Tao Yang
Nanchang Univ, Sch Informat Engn, Nanchang 330031, Jiangxi, Peoples R China
1 shared publicationYan-Qiu Guan
Nanchang Univ, Sch Informat Engn, Nanchang 330031, Jiangxi, Peoples R China
1 shared publication- Yi Chang
Huazhong University of Science and Technology
1 shared publication
Latest Publications
- 30 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 33 (10577149) , 69-81
An Efficient Single Image De-Raining Model With Decoupled Deep Networks
Single image de-raining is an emerging paradigm for many outdoor computer vision applications since rain streaks can significantly degrade the visibility and render the function compromised. The introduction of deep learning (DL) has brought about substantial advancement on de-raining methods. However, most existing DL-based methods use single homogeneous network architecture to generate de-rained images in a general image restoration manner, ignoring the discrepancy between rain location detection and rain intensity estimation. We find that this discrepancy would cause feature interference and representation ability degradation problems which significantly affect de-raining performance. In this paper, we propose a novel heterogeneous de-raining architecture aiming to decouple rain location detection and rain intensity estimation (DLINet). For these two subtasks, we provide dedicated network structures according to their differential properties to meet their respective performance requirements. To coordinate the decoupled subnetworks, we develop a high-order collaborative network learning the dynamic inter-layer interactions between rain location and intensity. To effectively supervise the decoupled subnetworks during training, we propose a novel training strategy that imposes task-oriented supervision using the label learned via joint training. Extensive experiments on synthetic datasets and real-world rainy scenes demonstrate that the proposed method has great advantages over existing state-of-the-art methods.
- 30 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 33 (10577149) , 69-81
An Efficient Single Image De-Raining Model With Decoupled Deep Networks
Single image de-raining is an emerging paradigm for many outdoor computer vision applications since rain streaks can significantly degrade the visibility and render the function compromised. The introduction of deep learning (DL) has brought about substantial advancement on de-raining methods. However, most existing DL-based methods use single homogeneous network architecture to generate de-rained images in a general image restoration manner, ignoring the discrepancy between rain location detection and rain intensity estimation. We find that this discrepancy would cause feature interference and representation ability degradation problems which significantly affect de-raining performance. In this paper, we propose a novel heterogeneous de-raining architecture aiming to decouple rain location detection and rain intensity estimation (DLINet). For these two subtasks, we provide dedicated network structures according to their differential properties to meet their respective performance requirements. To coordinate the decoupled subnetworks, we develop a high-order collaborative network learning the dynamic inter-layer interactions between rain location and intensity. To effectively supervise the decoupled subnetworks during training, we propose a novel training strategy that imposes task-oriented supervision using the label learned via joint training. Extensive experiments on synthetic datasets and real-world rainy scenes demonstrate that the proposed method has great advantages over existing state-of-the-art methods.
- 3 August 2020
- Published by Optica Publishing Group in Applied Optics
- Vol. 59 (23) , 6828-6837
Deep-learning-based single-photon-counting compressive imaging via jointly trained subpixel convolution sampling
The combination of single-pixel-imaging and single-photon-counting technology can achieve ultrahigh-sensitivity photon-counting imaging. However, its applications in high-resolution and real-time scenarios are limited by the long sampling and reconstruction time. Deep-learning-based compressive sensing provides an effective solution due to its ability to achieve fast and high-quality reconstruction. This paper proposes a sampling and reconstruction integrated neural network for single-photon-counting compressive imaging. To effectively remove the blocking artefact, a subpixel convolutional layer is jointly trained with a deep reconstruction network to imitate compressed sampling. By modifying the forward and backward propagation of the network, the first layer is trained into a binary matrix, which can be applied to the imaging system. An improved deep-reconstruction network based on the traditional Inception network is proposed, and the experimental results show that its reconstruction quality is better than existing deep-learning-based compressive sensing reconstruction algorithms.
- 3 August 2020
- Published by Optica Publishing Group in Applied Optics
- Vol. 59 (23) , 6828-6837
Deep-learning-based single-photon-counting compressive imaging via jointly trained subpixel convolution sampling
The combination of single-pixel-imaging and single-photon-counting technology can achieve ultrahigh-sensitivity photon-counting imaging. However, its applications in high-resolution and real-time scenarios are limited by the long sampling and reconstruction time. Deep-learning-based compressive sensing provides an effective solution due to its ability to achieve fast and high-quality reconstruction. This paper proposes a sampling and reconstruction integrated neural network for single-photon-counting compressive imaging. To effectively remove the blocking artefact, a subpixel convolutional layer is jointly trained with a deep reconstruction network to imitate compressed sampling. By modifying the forward and backward propagation of the network, the first layer is trained into a binary matrix, which can be applied to the imaging system. An improved deep-reconstruction network based on the traditional Inception network is proposed, and the experimental results show that its reconstruction quality is better than existing deep-learning-based compressive sensing reconstruction algorithms.
- 5 November 2009
- Published by Institution of Engineering and Technology (IET) in Electronics Letters
- Vol. 45 (23) , 1184-1185
Gegenbauer orthogonal basis neural network and its weights-direct-determination method
A feedforward neural-network is proposed with hidden-layer neurons activated by Gegenbauer orthogonal functions. Compared with conventional back-propagation training algorithms, a weights-direct-determination method is proposed to determine the neural-network weights directly, which avoids a lengthy iterative-training procedure. Computer simulations further substantiate the efficacy of such a Gegenbauer orthogonal basis neural network.
- 5 November 2009
- Published by Institution of Engineering and Technology (IET) in Electronics Letters
- Vol. 45 (23) , 1184-1185
Gegenbauer orthogonal basis neural network and its weights-direct-determination method
A feedforward neural-network is proposed with hidden-layer neurons activated by Gegenbauer orthogonal functions. Compared with conventional back-propagation training algorithms, a weights-direct-determination method is proposed to determine the neural-network weights directly, which avoids a lengthy iterative-training procedure. Computer simulations further substantiate the efficacy of such a Gegenbauer orthogonal basis neural network.