A5100738070- Video Surveillance and Tracking Methods
- Advanced Measurement and Detection Methods
- Visual perception and processing mechanisms
- Child and Animal Learning Development
- Advanced Vision and Imaging
- Action Observation and Synchronization
- Infrared Target Detection Methodologies
- Supply Chain and Inventory Management
- Metallurgy and Material Forming
- Industrial Vision Systems and Defect Detection
- Image Processing Techniques and Applications
- Advanced Algorithms and Applications
- Advanced Image Fusion Techniques
- Indoor and Outdoor Localization Technologies
- Metal Forming Simulation Techniques
- Quality and Supply Management
- Multisensory perception and integration
- Visual Attention and Saliency Detection
- Face Recognition and Perception
- Image Enhancement Techniques
- Flow Measurement and Analysis
- Remote Sensing and Land Use
- Image and Signal Denoising Methods
- Neural and Behavioral Psychology Studies
- IoT-based Smart Home Systems
University of Electronic Science and Technology of China
2011-2024
University of California, Los Angeles
2019-2023
Jiangsu University of Science and Technology
2023
Jiangxi Science and Technology Normal University
2021
Jiangxi Normal University
2021
Tianjin University of Technology
2018-2021
Air Force Engineering University
2019-2021
North China Electric Power University
2012-2020
Chang'an University
2017-2020
Shenzhen University
2019
Imagine a pack of predators stalking their prey. The may not always move directly toward target (e.g., when circling around it), but they be consistently facing it. human visual system appears to extremely sensitive such situations, even in displays involving simple shapes. We demonstrate this by introducing the wolfpack effect, which is found several randomly moving, oriented shapes (darts, or discs with "eyes") point moving disc. Despite randomness shapes' movement, seem interact disc--as...
Certain motion patterns can cause even simple geometric shapes to be perceived as animate. Viewing such displays evokes strong activation in temporoparietal cortex, including areas and near the (predominantly right) posterior superior temporal sulcus (pSTS). These brain regions are sensitive socially relevant information, but nature of social information represented pSTS is unclear. For example, previous studies have been unable explore perception shifting intentions beyond animacy. This due...
Visual experience involves not only physical features such as color and shape, but also higher-level properties animacy goal-directedness. Perceiving is an inherently dynamic experience, in part because agents' goal-directed behavior may be frequently flux-unlike many of their properties. How does the visual system maintain update representations animate over time motion? The present study explored this question context a particularly salient form perceived animacy: chasing, which one shape...
Mainstream theories of visual perception assume that working memory (VWM) is critical for integrating online perceptual information and constructing coherent experiences in changing environments. Given the dynamic interaction between VWM, we propose how processed during can directly determine going to be selected, consolidated, maintained VWM. We demonstrate validity this hypothesis by investigating what kinds stored as integrated objects Three criteria object-based storage are introduced:...
Two central tasks of visual processing are (1) to segment undifferentiated retinal images into discrete objects, and (2) represent those objects as the same persisting individuals over time motion. Here we explore interaction these two types in context object files—mid-level representations that "stick" moving on basis spatiotemporal properties. Object files can be revealed by object-specific preview benefits (OSPBs), wherein a "preview" information speeds recognition at later point when it...
Recent progress in deep learning is essentially based on a "big data for small tasks" paradigm, under which massive amounts of are used to train classifier single narrow task. In this paper, we call shift that flips paradigm upside down. Specifically, propose "small big wherein artificial intelligence (AI) system challenged develop "common sense," enabling it solve wide range tasks with little training data. We illustrate the potential power new by reviewing models common sense synthesize...
The interspinous process device (IPD) has emerged as a viable alternative for managing lumbar degenerative pathologies. Nevertheless, limited research exists regarding mechanical failure modes including and spinous fracture. This study developed novel IPD (IPD-NEW) systematically evaluated its biomechanical characteristics through finite element (FE) analysis in vitro cadaveric testing. Six human L1–L5 specimens were subjected to testing under four experimental conditions: (1) Intact spine...
The right posterior superior temporal sulcus (pSTS) is a neural region involved in assessing the goals and intentions underlying motion of social agents. Recent research has identified visual cues, such as chasing, that trigger animacy detection intention attribution. When readily available display, these cues reliably activate pSTS. Here, using functional magnetic resonance imaging, we examined if attributing to random would likewise engage Participants viewed displays four moving circles...
We study the machine's understanding of embodied reference: One agent uses both language and gesture to refer an object another in a shared physical environment. Of note, this new visual task requires multimodal cues with perspective-taking identify which is being referred to. To tackle problem, we introduce YouRefIt, crowd-sourced dataset reference collected various scenes; contains 4,195 unique clips 432 indoor scenes. best our knowledge, first that allows us referring expressions daily...
Abstract Human vision supports social perception by efficiently detecting agents and extracting rich information about their actions, goals, intentions. Here, we explore the cognitive architecture of perceived animacy constructing Bayesian models that integrate domain‐specific hypotheses agency with domain‐general constraints on sensory, memory, attentional processing. Our model posits combines a bottom–up, feature‐based, parallel search for goal‐directed movements top–down selection process...
For particle filtering tracking method, choosing was random to some degree according the dynamics equation, which may cause inaccurate results. To compensate, an improved method presented. A moving vehicle detected by redundant discrete wavelet transforms (RDWT), and then key points were obtained scale invariant feature transform. The matching in follow-up frames SIFT used as initial particles improve performance. Experimental results show that more centralize region of motion area presented...
Abstract Conventional manipulators with rigid structures and stiffness actuators have poor flexibility, limited obstacle avoidance capability, constrained workspace. Some developed flexible or soft in recent years the characteristics of infinite degrees freedom, high environmental adaptability, extended manipulation capability. However, these existing still cannot achieve shrinking motion independent control specified segments like animals, which hinders their applications. In this paper, a...
In recent decades, Building Information Modeling (BIM) is well-known and has been widely accepted by AEC industry since it could make less mistakes, conflicts, bring more profits as well improve the level of communication coordination. However, in Operation Maintenance phase, most owners including transit agencies, still lack knowledge about what BIM deliverables processes to demand how facilitate encourage other stakeholders apply technology. Therefore, this research aims support Public...
Moving vehicle detection and tracking is the key technology in intelligent traffic monitoring system. For shortcomings deficiencies of frame-subtraction method, a novel Marr wavelet, kernel-based background modeling method subtraction based on binary discrete wavelet transforms (BDWT) are introduced. The model keeps sample intensity values for each pixel image uses this to estimate probability density function intensity. estimated using new kernel estimation technique. current frame...
This paper presents a moving object detection method with background subtraction and shadow removal in the RGB color space. For subtraction, metrically trimmed mean is employed as robust estimate of model absolute deviation (MAD) adopted scale estimate. A cascading chromaticity difference estimator, brightness spatial analysis are explored to discriminate object. Our experimental results validate good performance proposed method.