Jinshan Su
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jinshan Su has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Haibo Zhou
School of Electronic Science and Engineering, Nanjing University, Nanjing, China
2 shared publications - Doncheng Yuan
School of Electronic Science and Engineering Nanjing University 210023 Nanjing China
1 shared publication - Donghua Zhou
Key Laboratory of Vibration Signal Capture and Intelligent Processing, School of Electronic Engineering, Yili Normal University, Yining 835000, China
1 shared publication Haibo Zhou
Nanjing University
1 shared publicationHanming Li
Yili Normal University
1 shared publication- Hanming Li
Key Laboratory of Vibration Signal Capture and Intelligent Processing, Yili Normal University, Yining, China
1 shared publication
Latest Publications
- Open Access
Fast Phase Calibration of Reconfigurable MZI Optical Processors via BFGS Quasi-Newton Optimization
Manufacturing errors introduce phase deviations in Mach–Zehnder interferometers (MZIs) that degrade the fidelity of optical processors. To address this issue, we employ a Broyden–Fletcher–Goldfarb–Shanno (BFGS) quasi-Newton method for phase calibration of a reconfigurable MZI optical processor based on the Reck architecture. By optimizing the mapping from the target matrix to the optical network, the method determines the optimized phase parameters of 12 phase shifters. Thermo-optic simulations are further used to establish the relationship between the applied bias voltage and the induced phase shift, providing a link between the optimized phase parameters and the electrical driving conditions. Compared with Particle Swarm Optimization (PSO), Genetic Algorithms (GA), and Gradient Descent with Momentum (GDM), the BFGS method provides faster convergence and high calibration fidelity. These results demonstrate an efficient approach for phase calibration of programmable MZI optical processors.Open AccessFast Phase Calibration of Reconfigurable MZI Optical Processors via BFGS Quasi-Newton Optimization
Manufacturing errors introduce phase deviations in Mach–Zehnder interferometers (MZIs) that degrade the fidelity of optical processors. To address this issue, we employ a Broyden–Fletcher–Goldfarb–Shanno (BFGS) quasi-Newton method for phase calibration of a reconfigurable MZI optical processor based on the Reck architecture. By optimizing the mapping from the target matrix to the optical network, the method determines the optimized phase parameters of 12 phase shifters. Thermo-optic simulations are further used to establish the relationship between the applied bias voltage and the induced phase shift, providing a link between the optimized phase parameters and the electrical driving conditions. Compared with Particle Swarm Optimization (PSO), Genetic Algorithms (GA), and Gradient Descent with Momentum (GDM), the BFGS method provides faster convergence and high calibration fidelity. These results demonstrate an efficient approach for phase calibration of programmable MZI optical processors. - Open Access
ST-Mamba: Spatial-Temporal Mamba for Traffic Flow Estimation Recovery using Limited Data
Traffic flow estimation (TFE) is crucial for urban intelligent traffic systems. While traditional on-road detectors are hindered by limited coverage and high costs, cloud computing and data mining of vehicular network data, such as driving speeds and GPS coordinates, present a promising and cost-effective alternative. Furthermore, minimizing data collection can significantly reduce overhead. However, limited data can lead to inaccuracies and instability in TFE. To address this, we introduce the spatial-temporal Mamba (ST-Mamba), a deep learning model combining a convolutional neural network (CNN) with a Mamba framework. ST-Mamba is designed to enhance TFE accuracy and stability by effectively capturing the spatial-temporal patterns within traffic flow. Our model aims to achieve results comparable to those from extensive data sets while only utilizing minimal data. Simulations using real-world datasets have validated our model's ability to deliver precise and stable TFE across an urban landscape based on limited data, establishing a cost-efficient solution for TFE.
Open AccessST-Mamba: Spatial-Temporal Mamba for Traffic Flow Estimation Recovery using Limited Data
Traffic flow estimation (TFE) is crucial for urban intelligent traffic systems. While traditional on-road detectors are hindered by limited coverage and high costs, cloud computing and data mining of vehicular network data, such as driving speeds and GPS coordinates, present a promising and cost-effective alternative. Furthermore, minimizing data collection can significantly reduce overhead. However, limited data can lead to inaccuracies and instability in TFE. To address this, we introduce the spatial-temporal Mamba (ST-Mamba), a deep learning model combining a convolutional neural network (CNN) with a Mamba framework. ST-Mamba is designed to enhance TFE accuracy and stability by effectively capturing the spatial-temporal patterns within traffic flow. Our model aims to achieve results comparable to those from extensive data sets while only utilizing minimal data. Simulations using real-world datasets have validated our model's ability to deliver precise and stable TFE across an urban landscape based on limited data, establishing a cost-efficient solution for TFE.
- 13 March 2024
- Published by Springer Nature in Peer-to-Peer Networking and Applications
- Vol. 17 (3) , 1667-1681
A UAV aided lightweight target information collection and detection approach
- 13 March 2024
- Published by Springer Nature in Peer-to-Peer Networking and Applications
- Vol. 17 (3) , 1667-1681
A UAV aided lightweight target information collection and detection approach
- Open Access
- 6 October 2023
- Published by Springer Nature in Research Square
A UAV Aided Lightweight Target Information Collection and Detection Approach
With the resumption of the World Cup and various concerts, the number of heavily crowded scenarios grows intensely. In these cases, it poses new challenges to massive information collection and lightweight target detection. Fortunately, the booming development of unmanned aerial vehicle (UAV) technology provides a highly flexible and cost-effective solution in many scenarios. This paper proposes a UAV aided lightweight target information collection and detection approach, where the target information is carried to a terrestrial distributed platform by a UAV and then a fast target detection is implemented. Firstly, we implement a 3D trajectory optimization for the UAV by minimizing the information collection time. Secondly, we design a lightweight target detection algorithm based on UAV loadable ARM (Advanced RISC Machines) architecture edge computing device. Finally, a terrestrial distributed processing platform is established. To ensure the stability and reliability of the target detection system, each module is tested separately. Numerical simulations show that, with the target detection module deployed on the Jetson Xavier NX edge computing platform for testing, the proposed target detection approach can achieve a detection accuracy of 89.5$\%$ and 71FPS detection speed using GPU acceleration compared with state-of-the-art methods.
- 6 October 2023
- Published by Springer Nature in Research Square
Open AccessA UAV Aided Lightweight Target Information Collection and Detection Approach
With the resumption of the World Cup and various concerts, the number of heavily crowded scenarios grows intensely. In these cases, it poses new challenges to massive information collection and lightweight target detection. Fortunately, the booming development of unmanned aerial vehicle (UAV) technology provides a highly flexible and cost-effective solution in many scenarios. This paper proposes a UAV aided lightweight target information collection and detection approach, where the target information is carried to a terrestrial distributed platform by a UAV and then a fast target detection is implemented. Firstly, we implement a 3D trajectory optimization for the UAV by minimizing the information collection time. Secondly, we design a lightweight target detection algorithm based on UAV loadable ARM (Advanced RISC Machines) architecture edge computing device. Finally, a terrestrial distributed processing platform is established. To ensure the stability and reliability of the target detection system, each module is tested separately. Numerical simulations show that, with the target detection module deployed on the Jetson Xavier NX edge computing platform for testing, the proposed target detection approach can achieve a detection accuracy of 89.5$\%$ and 71FPS detection speed using GPU acceleration compared with state-of-the-art methods.
- Open Access
- 1 August 2021
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1992 (4) , 042037
Intricate Face Recognition Based On Virtual Sample Generation
This paper mainly focuses on the problem of low recognition rate and making deep hidden features unable to learn caused by the lack of training samples in intricate facial recognition, a method of virtual image sample generation is proposed. The starGAN network based on style migration is used to generate the image that does not exist in the original database, and multiple images with different facial attributes are added to a single sample, combine the obtained virtual samples with the original database to build a larger face database for face recognition. Experiments on the CelebA face database show that the proposed method can effectively improve the face recognition rate in complex situations.
- 1 August 2021
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1992 (4) , 042037
Open AccessIntricate Face Recognition Based On Virtual Sample Generation
This paper mainly focuses on the problem of low recognition rate and making deep hidden features unable to learn caused by the lack of training samples in intricate facial recognition, a method of virtual image sample generation is proposed. The starGAN network based on style migration is used to generate the image that does not exist in the original database, and multiple images with different facial attributes are added to a single sample, combine the obtained virtual samples with the original database to build a larger face database for face recognition. Experiments on the CelebA face database show that the proposed method can effectively improve the face recognition rate in complex situations.