A5041779682- Advanced Neural Network Applications
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Intracerebral and Subarachnoid Hemorrhage Research
- Acute Ischemic Stroke Management
- MRI in cancer diagnosis
- Gear and Bearing Dynamics Analysis
- Antimicrobial Peptides and Activities
- Medical Image Segmentation Techniques
- Oral microbiology and periodontitis research
- Adhesion, Friction, and Surface Interactions
- Advanced Neuroimaging Techniques and Applications
- Tribology and Lubrication Engineering
- Face recognition and analysis
- Topic Modeling
- Psychology of Moral and Emotional Judgment
- Gaze Tracking and Assistive Technology
- Software Engineering Research
- Social and Intergroup Psychology
- Technology Use by Older Adults
- Medical Imaging and Analysis
- Machine Learning in Healthcare
- Dementia and Cognitive Impairment Research
- Deception detection and forensic psychology
- Frailty in Older Adults
Chinese University of Hong Kong
2023-2024
Northeastern University
2024
Chongqing Normal University
2022-2023
Carnegie Mellon University
2022-2023
Army Medical University
2023
University of Chinese Academy of Sciences
2020-2021
Huazhong Agricultural University
2018
Visual object detection has achieved unprecedented advance with the rise of deep convolutional neural networks. However, detecting tiny objects (for example persons less than 20 pixels) in large-scale images remains not well investigated. The extremely small raise a grand challenge about feature representation while massive and complex backgrounds aggregate risk false alarms. In this paper, we introduce new benchmark, referred to as TinyPerson, opening up promising direction for long...
FPN-based detectors have made significant progress in general object detection, e.g., MS COCO and PASCAL VOC. However, these fail certain application scenarios, tiny detection. In this paper, we argue that the top-down connections between adjacent layers FPN bring two-side influences for not only positive. We propose a novel concept, fusion factor, to control information deep deliver shallow layers, adapting After series of experiments analysis, explore how estimate an effective value factor...
Makeup and cosmetics offer the potential for self-expression reshaping of social roles visually impaired people. However, there exist barriers to conducting a beauty regime because reliance on visual information color variances in makeup. We present content analysis 145 YouTube videos demonstrate individuals' unique practices before, during, after doing Based makeup practices, we then conducted semi-structured interviews with 12 people discuss their perceptions challenges process more depth....
Detecting tiny objects (e.g., less than 20 × pixels) in large-scale images is an important yet open problem. Modern CNN-based detectors are challenged by the scale mismatch between dataset for network pre-training and target detector training. In this paper, we investigate alignment datasets, propose a new refined Scale Match method (termed SM+) person detection. SM+ improves match from image level to instance level, effectively promotes similarity dataset. Moreover, considering possibly...
Background: The changes of circadian rhythm are very common in senior adults, which can lead to decreased quality life and jeopardize the brain health long run. Among all types sleep disturbances, poor quality, as a modifiable condition, greatly affect daily activities, cognitive functions, life, even be implicated key potential contributing factor development accelerated decline prodromal dementia. Consequently, research efforts understand, thus potentially model, effects on cognition...
SPLUNC1 (short palate lung and nasal epithelial clone 1) is a multifunctional host defense protein found in human respiratory tract with antimicrobial properties. In this work, we compare the biological activities of four peptide (AMP) derivatives using paired clinical isolates Gram-negative (G(−)) bacteria Klebsiella pneumoniae, obtained from 11 patients with/without colistin resistance. Secondary structural studies were carried out to study interactions between AMPs lipid model membranes...
Background Poor sleep quality is increasingly considered to be an underlying cause of cerebrovascular diseases. This a slowly progressing condition that gradually leads vascular cognitive impairment and stroke during ageing. At present, randomized clinical trials examining the non-pharmacological therapies in management this comorbidity are very limited. Transcranial current stimulation (tCS) non-invasive technology for promoting function treating brain disorders. As advanced modalities tCS,...
The analysis of rough surface morphology plays an important role in the functional characteristics contact mechanical parts. Fractal geometry method is more accurate and sensitive than classical statistics model. For fractal representation surface, it necessary to determine proper dimension calculation method. In this research, effect power spectral density studied based on Monte-Carlo dimensions are calculated, theoretical calculated values compared with paired samples. And results by...
The 1st Tiny Object Detection (TOD) Challenge aims to encourage research in developing novel and accurate methods for tiny object detection images which have wide views, with a current focus on person detection. TinyPerson dataset was used the TOD is publicly released. It has 1610 72651 box-levelannotations. Around 36 participating teams from globe competed inthe Challenge. In this paper, we provide brief summary of the1st including introductions top three methods.The submission leaderboard...
Group interaction is an essential way of social and plays important role in our development. It has been found that when individuals participate group interactions, the identity partner affects mental processing behavioral decision-making subjects. However, little known about how deaf college students, who are labeled distinctly different from normal hearing will react facing proposers groups ultimatum game (UG) its time course. In this study, we recruited 29 students played UG which they...
Aneurysmal subarachnoid hemorrhage is a medical emergency of brain that has high mortality and poor prognosis. Causal effect estimation treatment strategies on patient outcomes crucial for aneurysmal decision-making. However, most existing studies decision-making support this disease are unable to simultaneously compare the potential different treatments patient. Furthermore, these fail harmoniously integrate imaging data with non-imaging clinical data, both which useful in scenarios. In...
Eye gaze that reveals human observational patterns has increasingly been incorporated into solutions for vision tasks. Despite recent explorations on leveraging to aid deep networks, few studies exploit as an efficient annotation approach medical image segmentation which typically entails heavy annotating costs. In this paper, we propose collect dense weak supervision with a scheme. To train gaze, multi-level framework trains multiple networks from discriminative attention, simulated set of...
Code search uses natural language queries to retrieve code snippets from a vast database, identifying those that are semantically similar the query. This enables developers reuse and enhance software development efficiency. Most existing algorithms focus on capturing semantic structural features by learning both text graph structures. However, these often struggle capture deeper within sources, leading lower accuracy in results. To address this issue, paper proposes novel semantics-enhanced...
The contact performance of various mechanical parts has important influence on the static, dynamic and motion response equipment. characterization rough surface topography is basis for study friction, wear deformation between bonding surfaces. By analyzing models characteristic parameters interface, we provide a systematic research idea model joint in future. For further details profile, G-W put forward. Then M-B perfects its lack scale independence. relationship actual area load derived...
Brain midline shift (MLS) is one of the most critical factors to be considered for clinical diagnosis and treatment decision-making intracranial hemorrhage. Existing computational methods on MLS quantification not only require intensive labeling in millimeter-level measurement but also suffer from poor performance due their dependence specific landmarks or simplified anatomical assumptions. In this paper, we propose a novel semi-supervised framework accurately measure scale head CT scans. We...
Intracerebral hemorrhage (ICH) is the second most common and deadliest form of stroke. Despite medical advances, predicting treat ment outcomes for ICH remains a challenge. This paper proposes novel prognostic model that utilizes both imaging tabular data to predict treatment outcome ICH. Our trained on observational collected from non-randomized controlled trials, providing reliable predictions success. Specifically, we propose employ variational autoencoder generate low-dimensional score,...
To investigate whether fairness decision among deaf college students is the same to normal hearing students, we recorded event-related potentials (ERPs) examine how brain affected when perform ultimatum gaming task. For behavioral outcomes, a main effect of proposal type was found for both and more unfair offer, lower acceptance rate have, which indicated that have unfairness aversion fair decision-making as at level. The ERPs results revealed there significant interaction between group...
Abstract To investigate whether fairness decision among deaf college students is the sameto normal hearing students, we recorded event-related potentials (ERPs) to examine how brain affected when perform ultimatum gaming task. For behavioral outcomes, a main effect of proposal type was found for both and more unfair offer, lower acceptance rate have, which indicated that have same unfairness aversion fair decision-making as at level. The ERPs results revealed there significant interaction...
Visual object detection has achieved unprecedented ad-vance with the rise of deep convolutional neural networks.However, detecting tiny objects (for example per-sons less than 20 pixels) in large-scale images remainsnot well investigated. The extremely small raisea grand challenge about feature representation while themassive and complex backgrounds aggregate risk offalse alarms. In this paper, we introduce a new benchmark,referred to as TinyPerson, opening up promising directionfor long...