Search

Your search keyword '"Wang, Hanqi"' showing total 252 results
Start Over Author "Wang, Hanqi"
252 results on '"Wang, Hanqi"'

1. Explain EEG-based End-to-end Deep Learning Models in the Frequency Domain

Author

Wang, Hanqi, Yang, Kun, Zhang, Jingyu, Chen, Tao, and Song, Liang

Subjects
Electrical Engineering and Systems Science - Signal Processing
Abstract

The recent rise of EEG-based end-to-end deep learning models presents a significant challenge in elucidating how these models process raw EEG signals and generate predictions in the frequency domain. This challenge limits the transparency and credibility of EEG-based end-to-end models, hindering their application in security-sensitive areas. To address this issue, we propose a mask perturbation method to explain the behavior of end-to-end models in the frequency domain. Considering the characteristics of EEG data, we introduce a target alignment loss to mitigate the out-of-distribution problem associated with perturbation operations. Additionally, we develop a perturbation generator to define perturbation generation in the frequency domain. Our explanation method is validated through experiments on multiple representative end-to-end deep learning models in the EEG decoding field, using an established EEG benchmark dataset. The results demonstrate the effectiveness and superiority of our method, and highlight its potential to advance research in EEG-based end-to-end models.

Published
2024

2. Cascaded Self-supervised Learning for Subject-independent EEG-based Emotion Recognition

Author

Wang, Hanqi, Chen, Tao, and Song, Liang

Subjects
Electrical Engineering and Systems Science - Signal Processing
Abstract

EEG-based Emotion recognition holds significant promise for applications in human-computer interaction, medicine, and neuroscience. While deep learning has shown potential in this field, current approaches usually rely on large-scale high-quality labeled datasets, limiting the performance of deep learning. Self-supervised learning offers a solution by automatically generating labels, but its inter-subject generalizability remains under-explored. For this reason, our interest lies in offering a self-supervised learning paradigm with better inter-subject generalizability. Inspired by recent efforts in combining low-level and high-level tasks in deep learning, we propose a cascaded self-supervised architecture for EEG emotion recognition. Then, we introduce a low-level task, time-to-frequency reconstruction (TFR). This task leverages the inherent time-frequency relationship in EEG signals. Our architecture integrates it with the high-level contrastive learning modules, performing self-supervised learning for EEG-based emotion recognition. Experiment on DEAP and DREAMER datasets demonstrates superior performance of our method over similar works. The outcome results also highlight the indispensability of the TFR task and the robustness of our method to label scarcity, validating the effectiveness of the proposed method.

Published
2024

3. Spatio-Temporal Domain Awareness for Multi-Agent Collaborative Perception

Author

Yang, Kun, Yang, Dingkang, Zhang, Jingyu, Li, Mingcheng, Liu, Yang, Liu, Jing, Wang, Hanqi, Sun, Peng, and Song, Liang

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Multi-agent collaborative perception as a potential application for vehicle-to-everything communication could significantly improve the perception performance of autonomous vehicles over single-agent perception. However, several challenges remain in achieving pragmatic information sharing in this emerging research. In this paper, we propose SCOPE, a novel collaborative perception framework that aggregates the spatio-temporal awareness characteristics across on-road agents in an end-to-end manner. Specifically, SCOPE has three distinct strengths: i) it considers effective semantic cues of the temporal context to enhance current representations of the target agent; ii) it aggregates perceptually critical spatial information from heterogeneous agents and overcomes localization errors via multi-scale feature interactions; iii) it integrates multi-source representations of the target agent based on their complementary contributions by an adaptive fusion paradigm. To thoroughly evaluate SCOPE, we consider both real-world and simulated scenarios of collaborative 3D object detection tasks on three datasets. Extensive experiments demonstrate the superiority of our approach and the necessity of the proposed components., Comment: Accepted by ICCV 2023

Published
2023

4. A novel efficient Multi-view traffic-related object detection framework

Author

Yang, Kun, Liu, Jing, Yang, Dingkang, Wang, Hanqi, Sun, Peng, Zhang, Yanni, Liu, Yan, and Song, Liang

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

With the rapid development of intelligent transportation system applications, a tremendous amount of multi-view video data has emerged to enhance vehicle perception. However, performing video analytics efficiently by exploiting the spatial-temporal redundancy from video data remains challenging. Accordingly, we propose a novel traffic-related framework named CEVAS to achieve efficient object detection using multi-view video data. Briefly, a fine-grained input filtering policy is introduced to produce a reasonable region of interest from the captured images. Also, we design a sharing object manager to manage the information of objects with spatial redundancy and share their results with other vehicles. We further derive a content-aware model selection policy to select detection methods adaptively. Experimental results show that our framework significantly reduces response latency while achieving the same detection accuracy as the state-of-the-art methods., Comment: Accepted by ICASSP 2023

Published
2023

5. Rethinking Saliency Map: An Context-aware Perturbation Method to Explain EEG-based Deep Learning Model

Author

Wang, Hanqi, Zhu, Xiaoguang, Chen, Tao, Li, Chengfang, and Song, Liang

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Electrical Engineering and Systems Science - Signal Processing
Abstract

Deep learning is widely used to decode the electroencephalogram (EEG) signal. However, there are few attempts to specifically investigate how to explain the EEG-based deep learning models. We conduct a review to summarize the existing works explaining the EEG-based deep learning model. Unfortunately, we find that there is no appropriate method to explain them. Based on the characteristic of EEG data, we suggest a context-aware perturbation method to generate a saliency map from the perspective of the raw EEG signal. Moreover, we also justify that the context information can be used to suppress the artifacts in the EEG-based deep learning model. In practice, some users might want a simple version of the explanation, which only indicates a few features as salient points. To this end, we propose an optional area limitation strategy to restrict the highlighted region. To validate our idea and make a comparison with the other methods, we select three representative EEG-based models to implement experiments on the emotional EEG dataset DEAP. The results of the experiments support the advantages of our method.

Published
2022

23. Stack Multiple Shallow Autoencoders into a Strong One: A New Reconstruction-Based Method to Detect Anomaly

Author

Wang, Hanqi, Hu, Xing, Song, Liang, Zhang, Guanhua, Liu, Yang, Liu, Jing, Jiang, Linhua, Goos, Gerhard, Founding Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Woeginger, Gerhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Mantoro, Teddy, editor, Lee, Minho, editor, Ayu, Media Anugerah, editor, Wong, Kok Wai, editor, and Hidayanto, Achmad Nizar, editor

Published
2021
Full Text
View/download PDF

25. Rectal adenocarcinoma: Ex vivo 9.4T MRI—correlation with histopathologic treatment response to neoadjuvant chemoradiotherapy.

Author

Li, Zhihui, Yuan, Yuan, Liu, Minglu, Bo, Tingting, Ma, Xiaolu, Wang, Hanqi, Chen, Chen, Shi, Xiaohui, Wang, Hao, Bai, Chenguang, Ni, Xiang, Shao, Chengwei, Lu, Yong, Lu, Jianping, and Shen, Fu

Abstract

Objectives: To determine the imaging details and diagnostic information of the treatment response to neoadjuvant chemoradiotherapy (nCRT) of rectal adenocarcinoma at 9.4T magnetic resonance imaging (MRI) by ex vivo. Methods: Fifteen cases with locally advanced rectal cancer (LARC) followed by radical surgery after nCRT between September 2022 and February 2023 were recruited. Resected specimens were fixed in a perfluoropolyether‐filled test tube and scanned with a 3.0T and 9.4T MRI system ex vivo. The residual tumor depth and MRI‐based tumor regression grade (TRG) were subjectively assessed and then compared with the pathological findings. Results: The ex vivo 9.4T T2WI without fat suppression clearly differentiated tumor tissue, fibrosis and normal rectal wall, which clearly corresponded to the pathologic tissues of the rectal specimens. The TRG could be accurately assessed on ex vivo 9.4T images in 13/15 specimens (86.7%), while in 11/15 specimens (73.3%) on ex vivo 3.0T images. Conclusion: Ex vivo 9.4T MR imaging clearly displayed the components of rectal wall and proved excellent diagnostic performance for evaluating the treatment response to nCRT, which allow radiologists to understand and then assess more accurately the TRG of LARC after nCRT. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

26. Comparing the Effect of Mechanical Loading on Deep and Superficial Cartilage Using Quantitative UTE MRI.

Author

Wang, Hanqi, Li, Zhihui, Li, Qing, Sommer, Stefan, Chen, Tongtong, Sun, Yao, Wei, Hongjiang, Yan, Fuhua, and Lu, Yong

Subjects
CARTILAGE, PHYSICAL activity, PEARSON correlation (Statistics), BODY mass index, BODY weight
Abstract

Background: The biomechanical properties of deep and superficial cartilage may be different, yet in vivo MRI validation is required. Purpose: To compare the effect of mechanical loading on deep and superficial cartilage in young healthy adults using ultrashort echo time (UTE)‐T2* mapping. Study Type: Prospective, intervention. Subjects: Thirty‐one healthy adults (54.8% females, median age = 23 years). Field Strength/Sequence: 3‐T, PD‐FS, and UTE sequences with four echo times (TEs = 0.1, 0.5, 2.8, and 4.0 msec; 0.6 mm isotropic spatial resolution) of the left knee, acquired before and after loading exercise. Assessment: Quantitative UTE‐T2* maps of the entire knee were generated using UTE images of four TEs. In deep and superficial cartilage of patella, medial and lateral femur, medial and lateral tibia cartilage (PC, MFC, LFC, MTC, and LTC), which were segmented manually, cartilage thickness and T2* values before and after loading were measured, extracted, taken averages of, and compared. Scan–rescan repeatability was evaluated. Body weight and body mass index (BMI) data were collected. Physical activity levels were evaluated using International Physical Activity Questionnaire. Statistical Tests: Paired sample t‐tests, paired Wilcoxon Mann–Whitney tests, Pearson and Spearman correlation analyses, Kruskal–Wallis tests with post‐hoc Bonferroni correction. A P‐value <0.05 was considered statistically significant. Results: The scan–rescan repeatability was good (RMSA‐CV < 10%). After exercise, deep cartilage exhibited no significant differences in cartilage thickness (PPC = 0.576, PMTC = 0.991, PMFC = 0.899, PLTC = 0.861, PLFC = 0.290) and T2* values (PPC = 0.914, PMTC = 0.780, PMFC = 0.754, PLTC = 0.327, PLFC = 0.811), which both significantly decreased in superficial PC, MFC, LFC, and MTC. The T2* values of superficial MTC and deep MFC were moderately correlated with higher body weight (ρ = 0.431) and lower BMI (ρ = −0.499), respectively. Data Conclusion: Deep and superficial cartilage may respond differently to mechanical loading as assessed by UTE‐T2*. Evidence Level: 2 Technical Efficacy: Stage 3 [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

40. Real‐world data on efficacy and safety of osimertinib in non‐small cell lung cancer patients with EGFR T790M mutation detected by first and repeat rebiopsy.

Author

Wang, Yuenan, Zhang, Huanhuan, Zou, Yuxia, Ren, Xueru, Wang, Hanqi, Bai, Rubing, Xu, Ke, Xu, Yehong, and Zhang, Zhihong

Subjects
NON-small-cell lung carcinoma, EPIDERMAL growth factor receptors, OSIMERTINIB, CANCER patients
Abstract

Background: Osimertinib could effectively target epidermal growth factor receptor (EGFR) T790M resistance mutations in non‐small cell lung cancer (NSCLC), indicating that rebiopsy may be particularly important. However, the clinical benefit of repeat rebiopsy in T790M‐negative patients with NSCLC detected by the first rebiopsy is still unclear, and data on the efficacy and safety of osimertinib in patients with NSCLC who are T790M‐positive patients on a repeat rebiopsy remain rare. Methods: We retrospectively collected the clinical data of advanced NSCLC patients with common EGFR mutation who were treated with 1/2‐generation (1/2G) EGFR–tyrosine kinase inhibitors (TKIs) in first‐line therapy in our center from January 2018 to December 2020. The detection rate of T790M by first and repeat rebiopsy was recorded, and we also analyzed the efficacy and safety of osimertinib for T790M‐positive patients. Results: Among 190 common EGFR‐mutant patients who received 1/2G EGFR–TKIs with advanced NSCLC in the first‐line treatment, 141 patients developed progressive disease. In total, 110 of 141 accepted the first rebiopsy, with a T790M prevalence of 50.9% (56/110). In total, 43 T790M‐positive patients who received osimertinib were included in first rebiopsy group. Of 54 T790M‐negative patients detected by the first rebiopsy, 28 underwent repeated rebiopsy in subsequent clinical treatment, and 10 (35.7%) T790M‐positive cases were confirmed. In total, eight T790M‐positive patients treated with osimertinib were included in repeat rebiopsy group. Overall, 66 (60%) of 110 patients acquired a T790M mutation. In patients with the T790M mutation discovered by the first and repeat rebiopsy, osimertinib resulted in median progression‐free survival of 7 (95% confidence interval [CI]: 5.3–8.7) and 6 (95% CI: 4.7–7.3) months, respectively (p =.656). The median overall survival since osimertinib initiation for T790M‐positive patients at first rebiopsy was 20 (95% CI: 15.1–24.9) months and 19 (95% CI: 16.9–21.1) months, for those at repeated rebiopsy (p =.888). Adverse events of grade 3 or higher were similar in the two groups (25.6% vs. 12.5%, p =.616). There was no treatment‐related death in the two groups. Conclusions: Repeat rebiopsy can increase the T790M mutation positivity rate. Osimertinib showed similar efficacy and safety in T790M‐positive patients whether detected by the first or repeat rebiopsy. [ABSTRACT FROM AUTHOR]

Published
2023
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results