A5101732544- Face and Expression Recognition
- Video Surveillance and Tracking Methods
- Advanced Image Processing Techniques
- Image and Signal Denoising Methods
- Advanced Image and Video Retrieval Techniques
- Advanced Image Fusion Techniques
- Image Processing Techniques and Applications
- Visual Attention and Saliency Detection
- Image Retrieval and Classification Techniques
- Advanced Measurement and Detection Methods
- Infrared Target Detection Methodologies
- Advanced Neural Network Applications
- Face recognition and analysis
- Robotics and Sensor-Based Localization
- Video Analysis and Summarization
- Geophysical Methods and Applications
- Image Enhancement Techniques
- Transportation Planning and Optimization
- Medical Imaging Techniques and Applications
- Remote Sensing and Land Use
- Anomaly Detection Techniques and Applications
- Olfactory and Sensory Function Studies
- Medical Image Segmentation Techniques
- Human Pose and Action Recognition
- Fire Detection and Safety Systems
Northeastern University
2019-2024
Liaoning Cancer Hospital & Institute
2024
Guangdong University of Technology
2023
Machine Science
2023
Chongqing Jiaotong University
2023
Beihang University
2010-2022
University of California, Irvine
2021
Zhejiang University
2018-2020
Beijing Institute of Technology
2013-2020
Nanjing University of Science and Technology
2007-2019
Existing enhancement methods are empirically expected to help the high-level end computer vision task: however, that is observed not always be case in practice. We focus on object or face detection poor visibility enhancements caused by bad weathers (haze, rain) and low light conditions. To provide a more thorough examination fair comparison, we introduce three benchmark sets collected real-world hazy, rainy, low-light conditions, respectively, with annotated objects/faces. launched UG <sup...
Convolutional neural networks (CNNs) have achieved great successes in face recognition, which unfortunately comes at the cost of massive computation and storage consumption. Many compact recognition are thus proposed to resolve this problem, triplet loss is effective further improve performance these models. However, it normally employs a fixed margin all samples, neglects informative similarity structures between different identities. In paper, we borrow idea knowledge distillation define...
Gesture recognition is an important human-computer interaction interface. This article introduces a novel hand gesture system based on Leap Motion gen.2. In this system, spatial fuzzy matching (SFM) algorithm first presented by and fusing information to construct fused dataset. For dynamic recognition, initial frame correction strategy SFM proposed fast initialize the trajectory of test with respect A notable feature that it can run ordinary laptops due small size dataset, which accelerates...
Locations of images have been widely used in many application scenarios for large geotagged image corpora. As to that are not geographically tagged, we estimate their locations with the help set by content-based retrieval. Bag-of-words representation has utilized widely. However, individual visual word-based retrieval approach is effective expressing salient relationships region. In this paper, present an location estimation multisaliency enhancement. We first extract region-of-interests...
How to encode a face is widely studied problem in both pattern recognition and psychology literatures. Many feature descriptors, Gabor feature, local binary (LBP), edge orientation histogram, have been proposed. In this paper, we give comprehensive study of these descriptors under the framework principal component analysis (PCA) followed by linear discriminant (LDA), compared on three different popular similarity measures two correspondence strategies: holistic local. Moreover, present new...
Saliency detection is widely used in many visual applications like image segmentation, object recognition and classification. In this paper, we will introduce a new method to detect salient objects natural images. The approach based on regional principal color contrast modal, which incorporates low-level medium-level cues. allows simple computation of features two categories spatial relationships saliency map, achieving higher F-measure rates. At the same time, present an interpolation...
The information acquisition and automatic processing technology based on visual surveillance sensors in intelligent transportation system (ITS) has become an important application field of computer vision technology. first step a traffic usually needs to correctly detect objects from videos classify them into different categories. In this paper, the improved spatiotemporal sample consistency algorithm (STSC) is proposed, enhance robustness background subtraction complex scenes. To address...
This paper proposes a novel local texture description method which defines six human visual perceptual characteristics and selects the minimal subset of relevant as well nonredundant features based on principal component analysis (PCA). We assign characteristics, were originally defined by Tamura et al., with definition metrics so that these measurements reflect perception each characteristic more precisely. Then, we propose PCA‐based feature selection exploiting structure components set to...
Real-time and accurate vehicle tracking by Cameras Surveillance can provide strong support for the acquisition application of important traffic parameters, which is basis condition evaluation reasonable command dispatch. To deal with difficult problems research in a complex environments, such as occlusion, sudden illumination change, similar target interference real-time tracking, measures are taken follows. Firstly, existing color local entropy particle filter method improved. The symmetry...
Recently, single image super-resolution (SISR) has been widely applied in the field of remote sensing processing and obtained remarkable performance. However, existing CNN-based methods are unable to exploit shallow visual characteristics at global receptive fields, which results limited perceptual capability these models. Furthermore, low-resolution inputs features contain abundant low-frequency information, weighed channels space equally, hence limiting representational ability CNNs. To...
Portable computing devices handling multi gigabit-per-second (Gb/s) data rates are anticipated to enter the wireless market with rise of 5G protocol in future. The recent allocation above-100-GHz band by Federal Communications Commission creates a great opportunity for developing networks distributed base station nodes targeting high rates. This ever-in-creasing need higher calls novel transceiver architectures that address fundamental shortcomings conventional designs and can achieve tens...
In the near-decade, Visual SLAM (Simultaneous Localization and Mapping) system is becoming more important for navigation of unmanned aerial vehicle (UAV) system, because it effective to replace positioning devices such as GPS (Global Position System) in indoor scenes. However, there are still challenges when camera working low-texture The visual algorithm based on a single feature difficult obtain enough features. accuracy robustness whole will be reduced or even cannot work properly....
In order to solve the problems of face features extraction and long time consuming in network training, we propose a facial expression recognition based on multi-level feature fusion structure. The structure consists following three parts: block; multi-granularity unit; global composed residual connection. We evaluated our proposed algorithm CK+ dataset FER2013 dataset, finally achieved 94.07% accuracy 65.4% respectively. Experimental results show that can effectively improve tasks.
With increasing traffic every day, most cities in the world are facing serious problems, such as accidents, congestion and air pollution. Despite recent improvement of urban infrastructure, reasonable light scheduling still plays an important role alleviating these problems. It is a great challenge to schedule huge number lights efficiently. To solve this problem, we propose Hybrid cellular swarm optimization method (HCSO) optimize lights. HCSO achieves efficient flexible scheduling, which...
Voice activity detection (VAD) is one of the most challenging problems in field speech signal processing. The statistical model based VADs have been widely studied recent literatures, which usually utilize hangover algorithms to prevent clipping weak tails. However, little attention has paid on initial consonants, and non-negligible onset errors might be incurred especially when SNR low. Since Mandarin syllables start with an algorithm proposed this paper improve performance VAD for...
The accuracy and efficiency of tea bud harvesting critically hinge upon the precision speed detection. To address this, we introduce YOLOv5-LNH model for detection in natural environments, building YOLOv5-S model's foundation. In pursuit heightened speed, eliminate large object head meticulously fine-tune parameters neck network, thereby significantly reducing redundant parameters. Our evaluation employs a real-world dataset. test outcomes unmistakably indicate that achieves remarkably...