Shouwei Gao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Shouwei Gao has a total of 95 co-authors. Here are the top co-authors by number of shared publications:
- Banghua Yang
School of Rehabilitation Science, Shanghai University of Traditional Chinese Medicine, Shanghai, China
8 shared publications Ding Xu
Shanghai Drug Rehabilitation Administration Bureau, Shanghai 200080, P. R. China
4 shared publications- Xinxing Xia
School of Mechatronic Engineering and Automation, Shanghai University, Shanghai, China
4 shared publications Xuelin Gu
School of Mechanical and Electrical Engineering and Automation, Shanghai University, Shanghai 200444, P. R. China
4 shared publications- Hao Zhou
School of Mechatronic Engineering and Automation, Shanghai University, Shanghai 201900, China
3 shared publications Lin Feng Yan
Department of Radiology & Functional and Molecular Imaging Key Lab of Shaanxi Province, Tangdu Hospital, Fourth Military Medical University, Xi’an, Shaanxi 710038, P. R. China
3 shared publications
Latest Publications
- Open Access
- 27 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 655-662
AwF-SNN: Fixed-Mask Continuity-Preserving Temporal State-Transition Compression for Spiking Neural Networks
- 27 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 655-662
Open AccessAwF-SNN: Fixed-Mask Continuity-Preserving Temporal State-Transition Compression for Spiking Neural Networks
- Open Access
- 31 December 2025
- Published by MDPI AG in Electronics
- Vol. 15 (1) , 189
SpikingDynamicMaskFormer: Enhancing Efficiency in Spiking Neural Networks with Dynamic Masking
Spiking Neural Networks (SNNs) offer promising low-power alternatives to conventional neural models but often incur considerable redundancy in parameters and computations. To address these inefficiencies, we propose SpikingDynamicMaskFormer (SDMFormer), a novel framework that integrates dynamic masking and lightweight position encoding into a spike-based Transformer backbone. Specifically, our Dynamic Mask Encoder Block adaptively suppresses ineffective spike channels by learning mask parameters, reducing parameter count to 37.93–42.69% of the original Spikformer. Simultaneously, a redesigned lightweight position embedding replaces resource-intensive relative position convolutions, further lowering complexity. Experiments on three neuromorphic vision datasets—DVS128, CIFAR10-DVS and N-Caltech101—demonstrate that SDMFormer cuts energy consumption by 42.79–50.13% relative to Spikformer while maintaining or slightly surpassing accuracy. Moreover, compared with recent leading works, SDMFormer achieves competitive accuracy with substantially fewer parameters and delivers higher inference efficiency, reaching up to 196.20 img/s on CIFAR10-DVS. These results highlight the efficacy of combining event-driven attention with structured pruning and parameter-efficient position encoding, indicating the potential of SDMFormer for resource-efficient SNN deployment in low-power applications.
- 31 December 2025
- Published by MDPI AG in Electronics
- Vol. 15 (1) , 189
Open AccessSpikingDynamicMaskFormer: Enhancing Efficiency in Spiking Neural Networks with Dynamic Masking
Spiking Neural Networks (SNNs) offer promising low-power alternatives to conventional neural models but often incur considerable redundancy in parameters and computations. To address these inefficiencies, we propose SpikingDynamicMaskFormer (SDMFormer), a novel framework that integrates dynamic masking and lightweight position encoding into a spike-based Transformer backbone. Specifically, our Dynamic Mask Encoder Block adaptively suppresses ineffective spike channels by learning mask parameters, reducing parameter count to 37.93–42.69% of the original Spikformer. Simultaneously, a redesigned lightweight position embedding replaces resource-intensive relative position convolutions, further lowering complexity. Experiments on three neuromorphic vision datasets—DVS128, CIFAR10-DVS and N-Caltech101—demonstrate that SDMFormer cuts energy consumption by 42.79–50.13% relative to Spikformer while maintaining or slightly surpassing accuracy. Moreover, compared with recent leading works, SDMFormer achieves competitive accuracy with substantially fewer parameters and delivers higher inference efficiency, reaching up to 196.20 img/s on CIFAR10-DVS. These results highlight the efficacy of combining event-driven attention with structured pruning and parameter-efficient position encoding, indicating the potential of SDMFormer for resource-efficient SNN deployment in low-power applications.
- Open Access
- 10 December 2025
- Published by MDPI AG in Electronics
- Vol. 14 (24) , 4860
ASFNOformer—A Superior Frequency Domain Token Mixer in Spiking Transformer
As the third generation of neural networks, Spiking Neural Networks (SNNs) simulate the event-driven processing mode of the brain, offering superior energy efficiency and biological interpretability compared to traditional deep learning. Combining the architectural strengths of Transformers with SNNs has recently demonstrated high accuracy and significant potential. SNNs process binary spikes and rich temporal information, resulting in lower computational complexity and making them particularly suitable for neuromorphic datasets. However, neuromorphic data typically involve dynamic edges and high-frequency pixel intensity changes. Capturing this frequency information is challenging for traditional spatial methods but is critical for event-driven vision. To address this, we investigate the integration of the Fast Fourier Transform (FFT) into SNNs and propose the Adaptive Spiking Fourier Neural Operator Transformer (ASFNOformer). This architecture adapts the Adaptive Fourier Neural Operator (AFNO)—originally validated in Artificial Neural Networks (ANNs)—specifically for the spiking domain. Unlike standard AFNOs, our module applies FFT across both spatial (H, W) and temporal (T) dimensions, followed by a Multi-Layer Perceptron structure (MLP) mechanism with a block-diagonal weight matrix. This design effectively captures both spatial features and temporal dynamics inherent in event streams. Furthermore, we incorporate Leaky Integrate-and-Fire (LIF) neurons optimized with Learnable Weight Parameters (LWP-LIF) to enhance temporal feature extraction and adaptivity. Experimental results on standard benchmarks indicate that our method reduces the parameter count by approximately 25%. In terms of recognition accuracy, ASFNOformer is comparable to mainstream models on static datasets and demonstrates superior performance on neuromorphic datasets by efficiently capturing frequency features. Notably, ablation studies confirm the model’s generalizability, and when using QKformer as a baseline, our method achieves state-of-the-art (SOTA) performance on the CIFAR10-DVS dataset. This work advances frequency-domain analysis in SNNs, paving the way for efficient deployment on neuromorphic hardware.
- 10 December 2025
- Published by MDPI AG in Electronics
- Vol. 14 (24) , 4860
Open AccessASFNOformer—A Superior Frequency Domain Token Mixer in Spiking Transformer
As the third generation of neural networks, Spiking Neural Networks (SNNs) simulate the event-driven processing mode of the brain, offering superior energy efficiency and biological interpretability compared to traditional deep learning. Combining the architectural strengths of Transformers with SNNs has recently demonstrated high accuracy and significant potential. SNNs process binary spikes and rich temporal information, resulting in lower computational complexity and making them particularly suitable for neuromorphic datasets. However, neuromorphic data typically involve dynamic edges and high-frequency pixel intensity changes. Capturing this frequency information is challenging for traditional spatial methods but is critical for event-driven vision. To address this, we investigate the integration of the Fast Fourier Transform (FFT) into SNNs and propose the Adaptive Spiking Fourier Neural Operator Transformer (ASFNOformer). This architecture adapts the Adaptive Fourier Neural Operator (AFNO)—originally validated in Artificial Neural Networks (ANNs)—specifically for the spiking domain. Unlike standard AFNOs, our module applies FFT across both spatial (H, W) and temporal (T) dimensions, followed by a Multi-Layer Perceptron structure (MLP) mechanism with a block-diagonal weight matrix. This design effectively captures both spatial features and temporal dynamics inherent in event streams. Furthermore, we incorporate Leaky Integrate-and-Fire (LIF) neurons optimized with Learnable Weight Parameters (LWP-LIF) to enhance temporal feature extraction and adaptivity. Experimental results on standard benchmarks indicate that our method reduces the parameter count by approximately 25%. In terms of recognition accuracy, ASFNOformer is comparable to mainstream models on static datasets and demonstrates superior performance on neuromorphic datasets by efficiently capturing frequency features. Notably, ablation studies confirm the model’s generalizability, and when using QKformer as a baseline, our method achieves state-of-the-art (SOTA) performance on the CIFAR10-DVS dataset. This work advances frequency-domain analysis in SNNs, paving the way for efficient deployment on neuromorphic hardware.
- 20 June 2025
- Published by Springer Nature in Applied Intelligence
- Vol. 55 (11) , 1-15
SFNS: Spatial-frequency image noise suppression for low-power industrial cone-beam computed tomography
- 20 June 2025
- Published by Springer Nature in Applied Intelligence
- Vol. 55 (11) , 1-15
SFNS: Spatial-frequency image noise suppression for low-power industrial cone-beam computed tomography
- Open Access
- 15 February 2025
- Published by MDPI AG in Electronics
- Vol. 14 (4) , 761
SATRN: Spiking Audio Tagging Robust Network
Audio tagging, as a fundamental task in acoustic signal processing, has demonstrated significant advances and broad applications in recent years. Spiking Neural Networks (SNNs), inspired by biological neural systems, exploit event-driven computing paradigms and temporal information processing, enabling superior energy efficiency. Despite the increasing adoption of SNNs, the potential of event-driven encoding mechanisms for audio tagging remains largely unexplored. This work presents a pioneering investigation into event-driven encoding strategies for SNN-based audio tagging. We propose the SATRN (Spiking Audio Tagging Robust Network), a novel architecture that integrates temporal–spatial attention mechanisms with membrane potential residual connections. The network employs a dual-stream structure combining global feature fusion and local feature extraction through inverted bottleneck blocks, specifically designed for efficient audio processing. Furthermore, we introduce an event-based encoding approach that enhances the resilience of Spiking Neural Networks to disturbances while maintaining performance. Our experimental results on the Urbansound8k and FSD50K datasets demonstrate that the SATRN achieves comparable performance to traditional Convolutional Neural Networks (CNNs) while requiring significantly less computation time and showing superior robustness against noise perturbations, making it particularly suitable for edge computing scenarios and real-time audio processing applications.
- 15 February 2025
- Published by MDPI AG in Electronics
- Vol. 14 (4) , 761
Open AccessSATRN: Spiking Audio Tagging Robust Network
Audio tagging, as a fundamental task in acoustic signal processing, has demonstrated significant advances and broad applications in recent years. Spiking Neural Networks (SNNs), inspired by biological neural systems, exploit event-driven computing paradigms and temporal information processing, enabling superior energy efficiency. Despite the increasing adoption of SNNs, the potential of event-driven encoding mechanisms for audio tagging remains largely unexplored. This work presents a pioneering investigation into event-driven encoding strategies for SNN-based audio tagging. We propose the SATRN (Spiking Audio Tagging Robust Network), a novel architecture that integrates temporal–spatial attention mechanisms with membrane potential residual connections. The network employs a dual-stream structure combining global feature fusion and local feature extraction through inverted bottleneck blocks, specifically designed for efficient audio processing. Furthermore, we introduce an event-based encoding approach that enhances the resilience of Spiking Neural Networks to disturbances while maintaining performance. Our experimental results on the Urbansound8k and FSD50K datasets demonstrate that the SATRN achieves comparable performance to traditional Convolutional Neural Networks (CNNs) while requiring significantly less computation time and showing superior robustness against noise perturbations, making it particularly suitable for edge computing scenarios and real-time audio processing applications.
- 9 January 2025
- Published by Springer Nature in Cognitive Neurodynamics
- Vol. 19 (1) , 1-20
Sg-snn: a self-organizing spiking neural network based on temporal information
Neurodynamic observations indicate that the cerebral cortex evolved by self-organizing into functional networks, These networks, or distributed clusters of regions, display various degrees of attention maps based on input. Traditionally, the study of network self-organization relies predominantly on static data, overlooking temporal information in dynamic neuromorphic data. This paper proposes Temporal Self-Organizing (TSO) method for neuromorphic data processing using a spiking neural network. The TSO method incorporates information from multiple time steps into the selection strategy of the Best Matching Unit (BMU) neurons. It enables the coupled BMUs to radiate the weight across the same layer of neurons, ultimately forming a hierarchical self-organizing topographic map of concern. Additionally, we simulate real neuronal dynamics, introduce a glial cell-mediated Glial-LIF (Leaky Integrate-and-fire) model, and adjust multiple levels of BMUs to optimize the attention topological map.Experiments demonstrate that the proposed Self-organizing Glial Spiking Neural Network (SG-SNN) can generate attention topographies for dynamic event data from coarse to fine. A heuristic method based on cognitive science effectively guides the network’s distribution of excitatory regions. Furthermore, the SG-SNN shows improved accuracy on three standard neuromorphic datasets: DVS128-Gesture, CIFAR10-DVS, and N-Caltech 101, with accuracy improvements of 0.3%, 2.4%, and 0.54% respectively. Notably, the recognition accuracy on the DVS128-Gesture dataset reaches 99.3%, achieving state-of-the-art (SOTA) performance.
- 9 January 2025
- Published by Springer Nature in Cognitive Neurodynamics
- Vol. 19 (1) , 1-20
Sg-snn: a self-organizing spiking neural network based on temporal information
Neurodynamic observations indicate that the cerebral cortex evolved by self-organizing into functional networks, These networks, or distributed clusters of regions, display various degrees of attention maps based on input. Traditionally, the study of network self-organization relies predominantly on static data, overlooking temporal information in dynamic neuromorphic data. This paper proposes Temporal Self-Organizing (TSO) method for neuromorphic data processing using a spiking neural network. The TSO method incorporates information from multiple time steps into the selection strategy of the Best Matching Unit (BMU) neurons. It enables the coupled BMUs to radiate the weight across the same layer of neurons, ultimately forming a hierarchical self-organizing topographic map of concern. Additionally, we simulate real neuronal dynamics, introduce a glial cell-mediated Glial-LIF (Leaky Integrate-and-fire) model, and adjust multiple levels of BMUs to optimize the attention topological map.Experiments demonstrate that the proposed Self-organizing Glial Spiking Neural Network (SG-SNN) can generate attention topographies for dynamic event data from coarse to fine. A heuristic method based on cognitive science effectively guides the network’s distribution of excitatory regions. Furthermore, the SG-SNN shows improved accuracy on three standard neuromorphic datasets: DVS128-Gesture, CIFAR10-DVS, and N-Caltech 101, with accuracy improvements of 0.3%, 2.4%, and 0.54% respectively. Notably, the recognition accuracy on the DVS128-Gesture dataset reaches 99.3%, achieving state-of-the-art (SOTA) performance.
- Open Access
- 26 December 2024
- Published by MDPI AG in Electronics
- Vol. 14 (1) , 43
SGSAFormer: Spike Gated Self-Attention Transformer and Temporal Attention
Spiking neural networks (SNNs), a neural network model structure inspired by the human brain, have emerged as a more energy-efficient deep learning paradigm due to their unique spike-based transmission and event-driven characteristics. Combining SNNs with the Transformer model significantly enhances SNNs’ performance while maintaining good energy efficiency. The gating mechanism, which dynamically adjusts input data and controls information flow, plays an important role in artificial neural networks (ANNs). Here, we introduce this gating mechanism into SNNs and propose a novel spike Transformer model, called SGSAFormer, based on the Spikformer network architecture. We introduce the Spike Gated Linear Unit (SGLU) module to improve the Multi-layer perceptron (MLP) module in SNNs by adding a gating mechanism to enhance the model’s expressive power. We also incorporate Spike Gated Self-Attention (SGSA) to strengthen the network’s attention mechanism, improving its ability to capture temporal information and dynamic processing. Additionally, we propose a Temporal Attention (TA) module, which selects new filters for the input data along the temporal dimension and can substantially reduce energy consumption with only a slight decrease in accuracy. To validate the effectiveness of our approach, we conducted extensive experiments on several neuromorphic datasets. Our model outperforms other state-of-the-art models in terms of performance.
- 26 December 2024
- Published by MDPI AG in Electronics
- Vol. 14 (1) , 43
Open AccessSGSAFormer: Spike Gated Self-Attention Transformer and Temporal Attention
Spiking neural networks (SNNs), a neural network model structure inspired by the human brain, have emerged as a more energy-efficient deep learning paradigm due to their unique spike-based transmission and event-driven characteristics. Combining SNNs with the Transformer model significantly enhances SNNs’ performance while maintaining good energy efficiency. The gating mechanism, which dynamically adjusts input data and controls information flow, plays an important role in artificial neural networks (ANNs). Here, we introduce this gating mechanism into SNNs and propose a novel spike Transformer model, called SGSAFormer, based on the Spikformer network architecture. We introduce the Spike Gated Linear Unit (SGLU) module to improve the Multi-layer perceptron (MLP) module in SNNs by adding a gating mechanism to enhance the model’s expressive power. We also incorporate Spike Gated Self-Attention (SGSA) to strengthen the network’s attention mechanism, improving its ability to capture temporal information and dynamic processing. Additionally, we propose a Temporal Attention (TA) module, which selects new filters for the input data along the temporal dimension and can substantially reduce energy consumption with only a slight decrease in accuracy. To validate the effectiveness of our approach, we conducted extensive experiments on several neuromorphic datasets. Our model outperforms other state-of-the-art models in terms of performance.
- Open Access
- 23 January 2024
- Published by IGI Global Scientific Publishing in International Journal on Semantic Web and Information Systems
- Vol. 20 (1) , 1-20
SSVEP-Enhanced Threat Detection and Its Impact on Image Segmentation
Selective attention, essential in discerning visual stimuli, enables the identification of threats such as snakes—a prime evolutionary influence on the human visual system. This phenomenon is encapsulated in snake detection theory (SDT), which posits that our ancestors' need to recognize these predators led to specialized perceptual abilities. This investigation utilizes steady-state visual evoked potentials (SSVEP) alongside the random image structure evolution technique, which systematically increases visual clarity through the interpolation of random noise, to probe the neural mechanisms underpinning selective attention, with a focus on serpentine forms. These findings underscore snakes' unique perceptual impact due to their curved forms and scaled textures, crucial for quick recognition—advancing image semantic segmentation and recognition tech.. This is particularly relevant for security and wildlife management, showcasing the evolutionary progression and cognitive prowess of the human visual apparatus.
- 23 January 2024
- Published by IGI Global Scientific Publishing in International Journal on Semantic Web and Information Systems
- Vol. 20 (1) , 1-20
Open AccessSSVEP-Enhanced Threat Detection and Its Impact on Image Segmentation
Selective attention, essential in discerning visual stimuli, enables the identification of threats such as snakes—a prime evolutionary influence on the human visual system. This phenomenon is encapsulated in snake detection theory (SDT), which posits that our ancestors' need to recognize these predators led to specialized perceptual abilities. This investigation utilizes steady-state visual evoked potentials (SSVEP) alongside the random image structure evolution technique, which systematically increases visual clarity through the interpolation of random noise, to probe the neural mechanisms underpinning selective attention, with a focus on serpentine forms. These findings underscore snakes' unique perceptual impact due to their curved forms and scaled textures, crucial for quick recognition—advancing image semantic segmentation and recognition tech.. This is particularly relevant for security and wildlife management, showcasing the evolutionary progression and cognitive prowess of the human visual apparatus.
- Open Access
- 27 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 120-126
A Wearable Brain-Computer Interface System for Fatigue Detection in Driving
Fatigue driving poses a significant threat to driver safety, leading to an increased risk of traffic accidents and potential harm to both life and property. Traditional fatigue detection methods, such as machine vision, are low effective due to factors such as lighting conditions and obstructions such as the presence of glasses. To address this issue and achieve more accurate fatigue detection, we design a Brain-computer interface (BCI) fatigue driving detection system based on electroencephalography (EEG) and electrooculography (EOG), specifically for measuring the fatigue level of drivers. The system includes three categories of fatigue level recognition. EEG and EOG data obtained from 10 subjects are classified using algorithms encapsulated in the main system. Remarkably, the system achieves an impressive average accuracy of 80.475% in recognizing fatigue levels, providing timely fatigue driving warnings as crucial feedback to the subjects. Additionally, the main system offers interactive features, including managing driver information and monitoring driver status. The system achieves real-time fatigue driving warnings and proposes a new method for detecting driving fatigue.
- 27 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 120-126
Open AccessA Wearable Brain-Computer Interface System for Fatigue Detection in Driving
Fatigue driving poses a significant threat to driver safety, leading to an increased risk of traffic accidents and potential harm to both life and property. Traditional fatigue detection methods, such as machine vision, are low effective due to factors such as lighting conditions and obstructions such as the presence of glasses. To address this issue and achieve more accurate fatigue detection, we design a Brain-computer interface (BCI) fatigue driving detection system based on electroencephalography (EEG) and electrooculography (EOG), specifically for measuring the fatigue level of drivers. The system includes three categories of fatigue level recognition. EEG and EOG data obtained from 10 subjects are classified using algorithms encapsulated in the main system. Remarkably, the system achieves an impressive average accuracy of 80.475% in recognizing fatigue levels, providing timely fatigue driving warnings as crucial feedback to the subjects. Additionally, the main system offers interactive features, including managing driver information and monitoring driver status. The system achieves real-time fatigue driving warnings and proposes a new method for detecting driving fatigue.
- Open Access
- 2 April 2023
- Published by MDPI AG in Healthcare
- Vol. 11 (7) , 1014
Effects of Background Music on Mental Fatigue in Steady-State Visually Evoked Potential-Based BCIs
As a widely used brain–computer interface (BCI) paradigm, steady-state visually evoked potential (SSVEP)-based BCIs have the advantages of high information transfer rates, high tolerance for artifacts, and robust performance across diverse users. However, the incidence of mental fatigue from prolonged, repetitive stimulation is a critical issue for SSVEP-based BCIs. Music is often used as a convenient, non-invasive means of relieving mental fatigue. This study investigates the compensatory effect of music on mental fatigue through the introduction of different modes of background music in long-duration, SSVEP-BCI tasks. Changes in electroencephalography power index, SSVEP amplitude, and signal-to-noise ratio were used to assess participants’ mental fatigue. The study’s results show that the introduction of exciting background music to the SSVEP-BCI task was effective in relieving participants’ mental fatigue. In addition, for continuous SSVEP-BCI tasks, a combination of musical modes that used soothing background music during the rest interval phase proved more effective in reducing users’ mental fatigue. This suggests that background music can provide a practical solution for long-duration SSVEP-based BCI implementation.
- 2 April 2023
- Published by MDPI AG in Healthcare
- Vol. 11 (7) , 1014
Open AccessEffects of Background Music on Mental Fatigue in Steady-State Visually Evoked Potential-Based BCIs
As a widely used brain–computer interface (BCI) paradigm, steady-state visually evoked potential (SSVEP)-based BCIs have the advantages of high information transfer rates, high tolerance for artifacts, and robust performance across diverse users. However, the incidence of mental fatigue from prolonged, repetitive stimulation is a critical issue for SSVEP-based BCIs. Music is often used as a convenient, non-invasive means of relieving mental fatigue. This study investigates the compensatory effect of music on mental fatigue through the introduction of different modes of background music in long-duration, SSVEP-BCI tasks. Changes in electroencephalography power index, SSVEP amplitude, and signal-to-noise ratio were used to assess participants’ mental fatigue. The study’s results show that the introduction of exciting background music to the SSVEP-BCI task was effective in relieving participants’ mental fatigue. In addition, for continuous SSVEP-BCI tasks, a combination of musical modes that used soothing background music during the rest interval phase proved more effective in reducing users’ mental fatigue. This suggests that background music can provide a practical solution for long-duration SSVEP-based BCI implementation.