Feiwei Qin

China National Petroleum Corporation (China)

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Yong Peng

Hangzhou Dianzi University

Changmiao Wang

Shenzhen Research Institute of Big Data

Ruiquan Ge

Hangzhou Dianzi University

Wanzeng Kong

Hangzhou Dianzi University

Yanli Shao

Zhejiang University

Jing Bai

University of Hong Kong

Yiyu Huang

Hangzhou Dianzi University

Zizhao Wu

Hangzhou Dianzi University

Feiping Nie

Northwestern Polytechnical University

Binfeng Zou

Hangzhou Dianzi University

Cooperative Institutions

Hangzhou Dianzi University

110

Zhejiang University

Chinese Academy of Sciences

Shanghai Jiao Tong University

Peking University

Tsinghua University

University of Chinese Academy of Sciences

Sichuan University

Central South University

University of Science and Technology of China

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

CEKD:Cross ensemble knowledge distillation for augmented fine-grained data

Applied Intelligence (2022)

Ke Zhang Jin Fan Shaoli Huang Yongliang Qiao Yu Xiaofeng

Ensemble Learning

Feature (linguistics)

10.1007/s10489-022-03355-0

Cite

Citations (1)

A Real-Time Semantic Segmentation Approach for Autonomous Driving Scenes

Journal of Computer-Aided Design & Computer Graphics (2021)

Feiwei Qin Xiyue Shen Yong Peng Yanli Shao Wenqiang Yuan

An important part of autonomous driving is the perception of the driving environment of the car, which has created a strong demand for high precision semantic segmentation algorithms that can be run in real time on low-power mobile devices. However, when analyzing the factors that affect the accuracy and speed of the semantic segmentation network, it can be found that in the structure of the previous semantic segmentation algorithm, spatial information and context features are difficult to take into account at the same time, and using two networks to obtain spatial information and context information separately will increase the amount of calculation and storage. Therefore, a new structure is proposed that divides the spatial path and context path from the network based on the residual structure, and a two-path real-time semantic segmentation network is designed based on this structure. The network contains a feature fusion module and an attention refinement module, which are used to realize the function of fusing the multi-scale features of two paths and optimizing the output results of context path. The network is based on the PyTorch framework and uses NVIDIA 1080Ti graphics cards for experiments. On the road scene data set Cityscapes, mIoU reached 78.8%, and the running speed reached 27.5 fps.

Spatial contextual awareness

Feature (linguistics)

10.3724/sp.j.1089.2021.18631

Cite

Citations (4)

VGNet: Multimodal Feature Extraction and Fusion Network for 3D CAD Model Retrieval

IEEE Transactions on Multimedia (2024)

Feiwei Qin Zhan Gao Meie Fang C. L. Philip Chen Ping Li

10.1109/tmm.2024.3521706

Cite

Citations (0)

SCUNet++: Swin-UNet and CNN Bottleneck Hybrid Architecture with Multi-Fusion Dense Skip Connection for Pulmonary Embolism CT Image Segmentation*

2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) (2024)

Yifei Chen Binfeng Zou Zhaoxin Guo Yiyu Huang Yifan Huang

Pulmonary embolism (PE) is a prevalent lung disease that can lead to right ventricular hypertrophy and failure in severe cases, ranking second in severity only to myocardial infarction and sudden death. Pulmonary artery CT angiography (CTPA) is a widely used diagnostic method for PE. However, PE detection presents challenges in clinical practice due to limitations in imaging technology. CTPA can produce noises similar to PE, making confirmation of its presence time-consuming and prone to overdiagnosis. Nevertheless, the traditional segmentation method of PE can not fully consider the hierarchical structure of features, local and global spatial features of PE CT images. In this paper, we propose an automatic PE segmentation method called SCUNet++ (Swin Conv UNet++). This method incorporates multiple fusion dense skip connections between the encoder and decoder, utilizing the Swin Transformer as the encoder. And fuses features of different scales in the decoder subnetwork to compensate for spatial information loss caused by the inevitable downsampling in Swin-UNet or other state-of-the-art methods, effectively solving the above problem. We provide a theoretical analysis of this method in detail and validate it on publicly available PE CT image datasets FUMPE and CAD-PE. The experimental results indicate that our proposed method achieved a Dice similarity coefficient (DSC) of 83.47% and a Hausdorff distance 95th percentile (HD95) of 3.83 on the FUMPE dataset, as well as a DSC of 83.42% and an HD95 of 5.10 on the CAD-PE dataset. These findings demonstrate that our method exhibits strong performance in PE segmentation tasks, potentially enhancing the accuracy of automatic segmentation of PE and providing a powerful diagnostic tool for clinical physicians. Our source code and new FUMPE dataset are available at https://github.com/JustlfC03/SCUNet-plusplus.

Hausdorff distance

10.1109/wacv57701.2024.00758

Cite

Citations (10)

ReliefNet: Fast Bas-relief Generation from 3D Scenes

Computer-Aided Design (2020)

Zhongping Ji Wei Feng Xianfang Sun Feiwei Qin Yigang Wang

10.1016/j.cad.2020.102928

Cite

Citations (24)

XLSTM-HVED: Cross-Modal Brain Tumor Segmentation and MRI Reconstruction Method Using Vision XLSTM and Heteromodal Variational Encoder-Decoder

arXiv (Cornell University) (2024)

Shenghao Zhu Yifei Chen Shuo Jiang Weihong Chen Chang Liu

Neurogliomas are among the most aggressive forms of cancer, presenting considerable challenges in both treatment and monitoring due to their unpredictable biological behavior. Magnetic resonance imaging (MRI) is currently the preferred method for diagnosing and monitoring gliomas. However, the lack of specific imaging techniques often compromises the accuracy of tumor segmentation during the imaging process. To address this issue, we introduce the XLSTM-HVED model. This model integrates a hetero-modal encoder-decoder framework with the Vision XLSTM module to reconstruct missing MRI modalities. By deeply fusing spatial and temporal features, it enhances tumor segmentation performance. The key innovation of our approach is the Self-Attention Variational Encoder (SAVE) module, which improves the integration of modal features. Additionally, it optimizes the interaction of features between segmentation and reconstruction tasks through the Squeeze-Fusion-Excitation Cross Awareness (SFECA) module. Our experiments using the BraTS 2024 dataset demonstrate that our model significantly outperforms existing advanced methods in handling cases where modalities are missing. Our source code is available at https://github.com/Quanato607/XLSTM-HVED.

10.48550/arxiv.2412.07804

Cite

Citations (0)

GFIL: A Unified Framework for the Importance Analysis of Features, Frequency Bands, and Channels in EEG-Based Emotion Recognition

IEEE Transactions on Cognitive and Developmental Systems (2021)

Yong Peng Feiwei Qin Wanzeng Kong Yuan Ge Feiping Nie

Accurately and automatically recognizing the emotional states of human beings is the central task in affective computing. The electroencephalography (EEG) data, generated from the neural activities in brain cortex, provide us with a reliable data source to perform emotion recognition. Besides the recognition accuracy, it is also necessary to explore the importance of different EEG features, frequency bands, and channels in emotion expression. In this article, we propose a unified framework termed graph-regularized least square regression with feature importance learning (GFIL) to simultaneously achieve these goals by incorporating an autoweighting variable into the least square regression. Unlike the widely used trial-and-error manner, GFIL automatically completes the identification once it is trained. Specifically, GFIL can: 1) adaptively discriminate the contributions of different feature dimensions; 2) automatically identify the critical frequency bands and channels; and 3) quantitatively rank and select the features by the learned autoweighting variable. From the experimental results on the SEED_IV data set, we find GFIL obtained improved accuracies based on the feature autoweighting strategy, which are 75.33%, 75.03%, and 79.17% corresponding to the three cross-session recognition tasks (session1->session2, session1->session3, session2->session3), respectively. Additionally, the Gamma band is identified as the most important one and the channels locating in the prefrontal and left/right central regions are more important.

Feature (linguistics)

10.1109/tcds.2021.3082803

Cite

Citations (59)

A Phoneme Sequence Driven Lightweight End-To-End Speech Synthesis Approach

Journal of Physics Conference Series (2019)

Zite Jiang Feiwei Qin Liaoying Zhao

Abstract This paper develops an end-to-end neural network model for text-to-speech (TTS) system based on phoneme sequence. Inspired by the Tacotron-2, the proposed model adopts an encoder-decoder model with attention mechanism and applies mel-spectrogram to measure the intermediate acoustic feature. Phoneme sequence is used to replace the character sequence in order to overcome the shortage of the character feature used in Tacotron-2. Unlike the conventional concatenate methodology based TTS system, our model can generate waveform directly from phoneme sequence. In addition, analogue from text analysis, a new analysis methodology is proposed for phoneme analysis. Experiment result on LJ Speech dataset shows that, compared with char-based model, our model can get a comparative or better performance.

Spectrogram

Sequence (biology)

Feature (linguistics)

Sequence labeling

End-to-end principle

10.1088/1742-6596/1267/1/012052

Cite

Citations (2)

VQ-NeRV: A Vector Quantized Neural Representation for Videos

arXiv (Cornell University) (2024)

Yunjie Xu Xiang Feng Feiwei Qin Ruiquan Ge Yong Peng

Implicit neural representations (INR) excel in encoding videos within neural networks, showcasing promise in computer vision tasks like video compression and denoising. INR-based approaches reconstruct video frames from content-agnostic embeddings, which hampers their efficacy in video frame regression and restricts their generalization ability for video interpolation. To address these deficiencies, Hybrid Neural Representation for Videos (HNeRV) was introduced with content-adaptive embeddings. Nevertheless, HNeRV's compression ratios remain relatively low, attributable to an oversight in leveraging the network's shallow features and inter-frame residual information. In this work, we introduce an advanced U-shaped architecture, Vector Quantized-NeRV (VQ-NeRV), which integrates a novel component--the VQ-NeRV Block. This block incorporates a codebook mechanism to discretize the network's shallow residual features and inter-frame residual information effectively. This approach proves particularly advantageous in video compression, as it results in smaller size compared to quantized features. Furthermore, we introduce an original codebook optimization technique, termed shallow codebook optimization, designed to refine the utility and efficiency of the codebook. The experimental evaluations indicate that VQ-NeRV outperforms HNeRV on video regression tasks, delivering superior reconstruction quality (with an increase of 1-2 dB in Peak Signal-to-Noise Ratio (PSNR)), better bit per pixel (bpp) efficiency, and improved video inpainting outcomes.

Representation

10.48550/arxiv.2403.12401

Cite

Citations (0)

An ontology-based semantic retrieval approach for heterogeneous 3D CAD models

Advanced Engineering Informatics (2016)

Feiwei Qin Shuming Gao Xiaoling Yang Ming Li Jing Bai

10.1016/j.aei.2016.10.001

Cite

Citations (44)