YCYuanpei Chen

Yuanpei Chen

*This is an algorithmically generated author record based on publicly available information.

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


Yuanpei Chen has a total of 37 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    ShipGen: Improving Remote Sensing Ship Detection via Layout-Conditioned Diffusion-Based Data Augmentation

    by Yuhan Zhang, Jun Liao, Jie Zhou, Yuanpei Chen, Xiaode Liu, Weihang Peng, Hongyu Yang, Qi Ming, Yufei Guo

    Synthetic Aperture Radar (SAR) ship detection plays a key role in the field of remote sensing. Data augmentation is frequently used to enhance detector performance due to the common challenge of limited annotated data. However, most augmented methods are limited to geometric transformations, “copy-paste”, or image inpainting, which are guided by strict constraints and suffer limited target and background diversity. To address this, we propose a novel multi-phase transfer-learned framework with layout-conditioned diffusion model. It incorporates sparse target layouts as relaxed conditions into the latent diffusion model, where the conditions take the form of bounding boxes that specify target positions and sizes. This leaves other ship attributes (e.g. orientation, scattering characteristics) and backgrounds to be freely generated and enhance target and background diversity. For enhancing the model's convergence on small datasets, we first pre-train it on large-scale datasets and then fine-tune it on small-scale ones for transfer learning. This enables precise layout-guided target generation even when using small datasets. For subsequent data augmentation, we adopt the widely used grounded-language-to-image generation (GLIGEN) model for practical use, and based on the transfer-learned model, we further develop a new automatic augmentation pipeline with a filtering mechanism. This pipeline integrates multiple foundation models and tools, including layout augmentation, segment anything model (SAM)-based annotation correction, and data filtering, and can automatically adjust inappropriate annotations and screen high-quality data to serve as augmented data for downstream detection tasks. Experiments on SSDD and HRSID datasets validate that the pipeline improves multiple detectors' performance across in-shore and off-shore scenarios, and qualitative analyses confirm high diversity and quality of generated images.


  • Nighttime Reflective Flare Removal From Images With Event Camera

    by Yuhan Zhang, Yuanpei Chen, Jie Zhou, Weihang Peng, Xiaode Liu, Zhe Ma, Yufei Guo

    As core imaging instruments in diverse measurement and monitoring scenarios, traditional frame-based cameras face a critical limitation in nighttime high-light environments, namely the inevitable emergence of reflective flares. Originating from repeated internal lens reflections of intense light sources, these flares accumulate over exposure time and appear as bright, irregular artifacts (e.g., “ghosting effects” or overexposed spots) in captured frames. Beyond visual defects, they introduce severe measurement errors that compromise the reliability of data from frame-based systems. By contrast, event cameras—emerging high-performance sensors—trigger signals asynchronously based on pixel-level light intensity changes. This unique capability lets them capture light source signals while minimizing reflective flares that are particularly noticeable at night, making them promising for providing prior information to remove flares from conventional camera images. To this end, we propose an innovative multi-modal fusion solution for nighttime reflective flare removal, integrating event streams with traditional images. Specifically, we introduce a simple yet effective event-guided fusion module that enhances events as prior knowledge to guide a subsequent multi-stage network in restoring flare-free images. Additionally, we contribute EventFlare, the first event-based reflective flare removal dataset with synthetic and real-world event-image pairs. Our experiments demonstrate the method’s effectiveness in efficiently removing reflective flares across both datasets.

  • Generative Model for 2.5D-Assisted Future Urban Remote Sensing Image Synthesis

    by Yuhan Zhang, Jie Zhou, Weihang Peng, Xiaode Liu, Yuanpei Chen

    Generating realistic future urban remote sensing imagery is critical for visualizing potential urban changes and supporting related technical analysis within urban planning. Traditional 2D-assisted methods are inherently limited in synthesize vertical development and infrastructure evolution, as they rely on binary planning maps. To address these limitations, we propose a novel 2.5D-assisted future urban remote sensing image synthesis method, aimed at generating future urban layouts based on existing urban structures and 2.5D planning maps. Specifically, the 2.5D map is divided into construction and demolition components, which are then integrated with the existing layout images and the embedding of the corresponding text as conditions for our generative model. We further design two trainable cascaded gated attention layers that process these two conditions separately and embed them into the latent diffusion model (LDM). This approach allows our model to dynamically comprehend the planning design requirements for key areas, making adjustments to accommodate diverse demands. Compared to existing state-of-the-art (SoTA) methods, our approach effectively targets design requirements, enabling flexible modifications that involve new constructions and demolitions in relevant urban areas. Experimental results on the 3DCD dataset demonstrate that the images generated by our method retain high fidelity and exhibit strong consistency with the 2.5D planning map.

  • Improved Event-Based Image De-Occlusion

    by Yufei Guo, Weihang Peng, Yuanpei Chen, Jie Zhou, Zhe Ma

    Reconstructing clear scene images in the presence of foreground occlusions remains a formidable challenge for cameras constrained by a single viewpoint. Synthetic aperture imaging (SAI) has emerged as a solution by integrating visual information from multiple viewpoints, thus overcoming occlusions. Event cameras, characterized by their high temporal resolution, exceptional dynamic range, low power consumption, and resistance to motion blur, offer a promising avenue for capturing intricate details of background objects within a limited range of motion. Leveraging these capabilities, several event-camera-based SAI methodologies have been proposed to effectively tackle dense occlusions. Despite these advancements, existing methodologies encounter obstacles stemming from the hybrid model they employ, comprising a spiking neural network (SNN) encoder and a convolutional neural network (CNN) decoder. Challenges include information degradation within the SNN encoder due to the quantization of full-precision data into binary spikes, as well as insufficient training data for the CNN decoder to adequately learn feature extraction. In response, we present an enhanced eventbased image de-occlusion approach. Our method introduces a novel full-precision leaky integrate-and-fire (FP-LIF) mechanism to mitigate information loss within the SNN encoder. Additionally, we propose an isomorphic network knowledge distillation method to augment the feature extraction capabilities of the CNN decoder. Experimental results demonstrate the efficacy of our approach in enhancing event-camera-based SAI methodologies.


  • Open Access

    Ternary Spike: Learning Ternary Spikes for Spiking Neural Networks

    by Yufei Guo, Yuanpei Chen, Xiaode Liu, Weihang Peng, Yuhan Zhang, Xuhui Huang, Zhe Ma

    The Spiking Neural Network (SNN), as one of the biologically inspired neural network infrastructures, has drawn increasing attention recently. It adopts binary spike activations to transmit information, thus the multiplications of activations and weights can be substituted by additions, which brings high energy efficiency. However, in the paper, we theoretically and experimentally prove that the binary spike activation map cannot carry enough information, thus causing information loss and resulting in accuracy decreasing. To handle the problem, we propose a ternary spike neuron to transmit information. The ternary spike neuron can also enjoy the event-driven and multiplication-free operation advantages of the binary spike neuron but will boost the information capacity. Furthermore, we also embed a trainable factor in the ternary spike neuron to learn the suitable spike amplitude, thus our SNN will adopt different spike amplitudes along layers, which can better suit the phenomenon that the membrane potential distributions are different along layers. To retain the efficiency of the vanilla ternary spike, the trainable ternary spike SNN will be converted to a standard one again via a re-parameterization technique in the inference. Extensive experiments with several popular network structures over static and dynamic datasets show that the ternary spike can consistently outperform state-of-the-art methods. Our code is open-sourced at https://github.com/yfguo91/Ternary-Spike.


  • Open Access

    NeuroCLIP: Neuromorphic Data Understanding by CLIP and SNN

    by Yufei Guo, Yuanpei Chen, Zhe Ma

    Recently, the neuromorphic vision sensor has received more and more interest. However, the neuromorphic data consists of asynchronous event spikes, which makes it difficult to construct a big benchmark to train a power general neural network model, thus limiting the neuromorphic data understanding for “unseen” objects by deep learning. While for the frame image, since the training data can be obtained easily, the zero-shot and few-shot learning for “unseen” task via the large Contrastive Vision-Language Pre-training (CLIP) model, which is pre-trained by large-scale image-text pairs in 2D, have shown inspirational performance. We wonder whether the CLIP could be transferred to neuromorphic data recognition to handle the “unseen” problem. To this end, we materialize this idea with NeuroCLIP in the paper. The NeuroCLIP consists of 2D CLIP and two specially designed modules for neuromorphic data understanding. First, an event-frame module that could convert the event spikes to the sequential frame image with a simple discrimination strategy. Second, an inter-timestep adapter, which is a simple fine-tuned adapter based on a spiking neural network (SNN) for the sequential features coming from the visual encoder of CLIP to improve the few-shot performance. Various experiments on neuromorphic datasets including N-MNIST, CIFAR10-DVS, and ES-ImageNet demonstrate the effectiveness of NeuroCLIP.


  • Open Access

    NeuroCLIP: Neuromorphic Data Understanding by CLIP and SNN

    by Yufei Guo, Yuanpei Chen, Zhe Ma

    Recently, the neuromorphic vision sensor has received more and more interest. However, the neuromorphic data consists of asynchronous event spikes, which makes it difficult to construct a big benchmark to train a power general neural network model, thus limiting the neuromorphic data understanding for ``unseen" objects by deep learning. While for the frame image, since the training data can be obtained easily, the zero-shot and few-shot learning for ``unseen" task via the large Contrastive Vision-Language Pre-training (CLIP) model, which is pre-trained by large-scale image-text pairs in 2D, have shown inspirational performance. We wonder whether the CLIP could be transferred to neuromorphic data recognition to handle the ``unseen" problem. To this end, we materialize this idea with NeuroCLIP in the paper. The NeuroCLIP consists of 2D CLIP and two specially designed modules for neuromorphic data understanding. First, an event-frame module that could convert the event spikes to the sequential frame image with a simple discrimination strategy. Second, an inter-timestep adapter, which is a simple fine-tuned adapter based on a spiking neural network (SNN) for the sequential features coming from the visual encoder of CLIP to improve the few-shot performance. Various experiments on neuromorphic datasets including N-MNIST, CIFAR10-DVS, and ES-ImageNet demonstrate the effectiveness of NeuroCLIP. Our code is open-sourced at https://github.com/yfguo91/NeuroCLIP.git.

  • Open Access

    InfLoR-SNN: Reducing Information Loss for Spiking Neural Networks

    by YuFei Guo, Yuanpei Chen, Liwen Zhang, YingLei Wang, Xiaode Liu, Xinyi Tong, Yuanyuan Ou, Xuhui Huang, Zhe Ma

    The Spiking Neural Network (SNN) has attracted more and more attention recently. It adopts binary spike signals to transmit information. Benefitting from the information passing paradigm of SNNs, the multiplications of activations and weights can be replaced by additions, which are more energy-efficient. However, its ``Hard Reset" mechanism for the firing activity would ignore the difference among membrane potentials when the membrane potential is above the firing threshold, causing information loss. Meanwhile, quantifying the membrane potential to 0/1 spikes at the firing instants will inevitably introduce the quantization error thus bringing about information loss too. To address these problems, we propose to use the ``Soft Reset" mechanism for the supervised training-based SNNs, which will drive the membrane potential to a dynamic reset potential according to its magnitude, and Membrane Potential Rectifier (MPR) to reduce the quantization error via redistributing the membrane potential to a range close to the spikes. Results show that the SNNs with the ``Soft Reset" mechanism and MPR outperform their vanilla counterparts on both static and dynamic datasets.


  • Open Access

    Ternary Spike: Learning Ternary Spikes for Spiking Neural Networks

    by Yufei Guo, Yuanpei Chen, Xiaode Liu, Weihang Peng, Yuhan Zhang, Xuhui Huang, Zhe Ma

    The Spiking Neural Network (SNN), as one of the biologically inspired neural network infrastructures, has drawn increasing attention recently. It adopts binary spike activations to transmit information, thus the multiplications of activations and weights can be substituted by additions, which brings high energy efficiency. However, in the paper, we theoretically and experimentally prove that the binary spike activation map cannot carry enough information, thus causing information loss and resulting in accuracy decreasing. To handle the problem, we propose a ternary spike neuron to transmit information. The ternary spike neuron can also enjoy the event-driven and multiplication-free operation advantages of the binary spike neuron but will boost the information capacity. Furthermore, we also embed a trainable factor in the ternary spike neuron to learn the suitable spike amplitude, thus our SNN will adopt different spike amplitudes along layers, which can better suit the phenomenon that the membrane potential distributions are different along layers. To retain the efficiency of the vanilla ternary spike, the trainable ternary spike SNN will be converted to a standard one again via a re-parameterization technique in the inference. Extensive experiments with several popular network structures over static and dynamic datasets show that the ternary spike can consistently outperform state-of-the-art methods. Our code is open-sourced at https://github.com/yfguo91/Ternary-Spike.

  • Open Access

    InfLoR-SNN: Reducing Information Loss for Spiking Neural Networks

    by Yufei Guo, Yuanpei Chen, Liwen Zhang, Xiaode Liu, Xinyi Tong, Yuanyuan Ou, Xuhui Huang, Zhe Ma

    The Spiking Neural Network (SNN) has attracted more and more attention recently. It adopts binary spike signals to transmit information. Benefitting from the information passing paradigm of SNNs, the multiplications of activations and weights can be replaced by additions, which are more energy-efficient. However, its "Hard Reset" mechanism for the firing activity would ignore the difference among membrane potentials when the membrane potential is above the firing threshold, causing information loss. Meanwhile, quantifying the membrane potential to 0/1 spikes at the firing instants will inevitably introduce the quantization error thus bringing about information loss too. To address these problems, we propose to use the "Soft Reset" mechanism for the supervised training-based SNNs, which will drive the membrane potential to a dynamic reset potential according to its magnitude, and Membrane Potential Rectifier (MPR) to reduce the quantization error via redistributing the membrane potential to a range close to the spikes. Results show that the SNNs with the "Soft Reset" mechanism and MPR outperform their vanilla counterparts on both static and dynamic datasets.