A5100456093- Photoreceptor and optogenetics research
- Emotion and Mood Recognition
- Advanced Memory and Neural Computing
- Advanced Neural Network Applications
- Speech and Audio Processing
- EEG and Brain-Computer Interfaces
- Image and Signal Denoising Methods
- Robotics and Sensor-Based Localization
- Video Analysis and Summarization
- Educational Practices and Challenges
- Video Surveillance and Tracking Methods
- Advanced Image Processing Techniques
- Neural Networks and Reservoir Computing
- Multimodal Machine Learning Applications
- Human Pose and Action Recognition
- Education and Critical Thinking Development
- Music and Audio Processing
- Perfectionism, Procrastination, Anxiety Studies
- Advanced Image and Video Retrieval Techniques
- Career Development and Diversity
- Anxiety, Depression, Psychometrics, Treatment, Cognitive Processes
- Ego Development and Educational Practices
- Mindfulness and Compassion Interventions
- Learning Styles and Cognitive Differences
- Advancements in Transdermal Drug Delivery
Srinakharinwirot University
2025
Nankai University
2022-2024
Yuxian People's Hospital
2024
Tianjin University
2019-2024
Xiamen University
2023
Chinese PLA General Hospital
2023
Beijing University of Posts and Telecommunications
2023
Qingdao University of Technology
2023
Peking University
2021-2022
Peking University Sixth Hospital
2021-2022
Abstract Bioinspired neuromorphic machine vision system (NMVS) that integrates retinomorphic sensing and computing into one monolithic is regarded as the most promising architecture for visual perception. However, large intensity range of natural lights complex illumination conditions in actual scenarios always require NMVS to dynamically adjust its sensitivity according environmental conditions, just like adaptation function human retina. Although some opto‐sensors with scotopic or photopic...
Abstract Artificial synapses and neurons are crucial milestones for neuromorphic computing hardware, memristors with resistive threshold switching characteristics regarded as the most promising candidates construction of hardware neural networks. However, can only operate in one mode, that is, or switching, distinct required to construct fully memristive making it more complex fabrication integration hardware. Herein, we propose a flexible dual‐mode memristor array based on core–shell CsPbBr...
Abstract Background Mindfulness-based cognitive therapy (MBCT) is a promising alternative treatment for generalized anxiety disorder (GAD). The objective of this study was to examine whether the efficacy group MBCT adapted treating GAD (MBCT-A) noninferior behavioural (CBT) designed treat (CBT-A), which considered one first-line treatments patients. We also explored MBCT-A in symptomatic patients compared with CBT-A variety outcomes symptoms, as well depressive overall illness severity,...
Automatically predicting the emotions of user-generated videos (UGVs) receives increasing interest recently. However, existing methods mainly focus on a few key visual frames, which may limit their capacity to encode context that depicts intended emotions. To tackle that, in this paper, we propose cross-modal temporal erasing network locates not only keyframes but also and audio-related information weakly-supervised manner. In specific, first leverage intra- inter-modal relationship among...
Existing methods in planar region segmentation suffer the problems of vague boundaries and failure to detect small-sized regions. To address these, this study presents an end-to-end framework, named PlaneSeg, which can be easily integrated into various plane models. Specifically, PlaneSeg contains three modules, namely, edge feature extraction module, multiscale resolution-adaptation module. First, module produces edge-aware maps for finer boundaries. The learned information acts as a...
Chronic stress has a negative impact on cognitive function and physical health, particularly impairing control function. This study aimed to investigate the effects of chronic error processing, specifically monitoring post-error adjustment. We assessed 61 healthy college students (32 females, 29 males, age: 18.80 ± 0.68 years) using Student-Life Stress Inventory (SLSI) four-choice Flanker task with response-stimulus intervals (RSIs) 200 ms, 700 1500 ms temporal dynamics characteristics...
Video sentiment analysis aims to uncover the underlying attitudes of viewers, which has a wide range applications in real world. Existing works simply classify video into single sentimental category, ignoring fact that untrimmed videos may appear multiple segments with varying lengths and unknown locations. To address this, we propose challenging task, i.e., Temporal Sentiment Localization (TSL), find parts convey sentiment. systematically investigate fully- weakly-supervised settings for...
Curcumin (CUR) is highly promising for topical therapeutic applications, but water-insolubility one of the major challenges plaguing its drugability, while conventional lipid nanocarriers are limited by low drug-carrying capacity, many additives, and complex processes. In current work, we constructed a composite carrier integrated with cyclodextrin metal-organic framework (γ-CD-MOF) nanosponge (β-CDNS), in which γ-CD-MOF had 13.9 % drug loading 267.1-fold increase solubility water CUR,...
Tracking both location and pose of multiple planar objects (MPOT) is great significance to numerous real-world applications. The greater degree-of-freedom compared with common makes MPOT far more challenging than well-studied object tracking, especially when occlusion occurs. To address this task, we are inspired by amodal perception that humans jointly track visible invisible parts the target, propose a tracking framework unifies appearance reasoning. Specifically, present dual-branch...
Recently, 2D convolution has been found unqualified in sound event detection (SED). It enforces translation equivariance on events along frequency axis, which is not a shift-invariant dimension. To address this issue, dynamic used to model the dependency of events. In paper, we proposed first full-dynamic method named \emph{full-frequency convolution} (FFDConv). FFDConv generates kernels for every band, designed directly structure frequency-dependent modeling. physically furnished with...
Camouflaged vision perception is an important task with numerous practical applications. Due to the expensive collection and labeling costs, this community struggles a major bottleneck that species category of its datasets limited small number object species. However, existing camouflaged generation methods require specifying background manually, thus failing extend sample diversity in low-cost manner. In paper, we propose Latent Background Knowledge Retrieval-Augmented Diffusion (LAKE-RED)...
Large Language Models (LLMs) excel in stand-alone code tasks like HumanEval and MBPP, but struggle with handling entire repositories. This challenge has prompted research on enhancing LLM-codebase interaction at a repository scale. Current solutions rely similarity-based retrieval or manual tools APIs, each notable drawbacks. Similarity-based often low recall complex tasks, while APIs are typically task-specific require expert knowledge, reducing their generalizability across diverse...
It is universally acknowledged that the financial trading a thorny issue in time-series scenarios.On one hand, due to great randomness and instability markets, existing machine learning methods are inadequate for modeling high-frequency data.On other it remains challenge identify validity of transaction actions avoid high fees.To address these issues, we propose novel framework, namely mTrader, offer suitable strategies automatically.We creatively design multiscale signal matrix describe...