A5100399485- Advanced Vision and Imaging
- Advanced Neural Network Applications
- Optical measurement and interference techniques
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Robotics and Sensor-Based Localization
- Image Processing Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- Image Enhancement Techniques
- Advanced Image Processing Techniques
- Generative Adversarial Networks and Image Synthesis
- Image Retrieval and Classification Techniques
- Sparse and Compressive Sensing Techniques
- Medical Image Segmentation Techniques
- Advanced Bandit Algorithms Research
- Statistical Methods and Inference
- Stochastic Gradient Optimization Techniques
- Anomaly Detection Techniques and Applications
- Brain Tumor Detection and Classification
- Industrial Vision Systems and Defect Detection
- Fire Detection and Safety Systems
- Human Pose and Action Recognition
- Advanced Algorithms and Applications
- Cancer-related molecular mechanisms research
- Remote Sensing and Land Use
Toronto Metropolitan University
2020-2025
Hunan Institute of Science and Technology
2009-2024
East China Normal University
2021-2024
Henan University
2022-2024
Guangzhou University
2023-2024
Shenzhen Second People's Hospital
2024
Hangzhou Normal University
2024
Chengdu University of Technology
2023
University of Toronto
2023
Zhengzhou University
2022-2023
Big Data analytics plays a key role through reducing the data size and complexity in applications. Visualization is an important approach to helping get complete view of discover values. visualization should be integrated seamlessly so that they work best Conventional methods as well extension some conventional applications are introduced this paper. The challenges discussed. New methods, applications, technology progress presented.
Learning depth from a single image, as an important issue in scene understanding, has attracted lot of attention the past decade. The accuracy estimation been improved conditional Markov random fields, non-parametric methods, to deep convolutional neural networks most recently. However, there exist inherent ambiguities recovering 3D 2D image. In this paper, we first prove ambiguity between focal length and monocular learning verify result using experiments, showing that great influence on...
This paper focuses on the problem of vision-based obstacle detection and tracking for unmanned aerial vehicle navigation. A real-time object localization strategy from monocular image sequences is developed by effectively integrating into a dynamic Kalman model. At stage, interest automatically detected localized saliency map computed via background connectivity cue at each frame; filter employed to provide coarse prediction state, which further refined local detector incorporating temporal...
Colonoscopy is a procedure to detect colorectal polyps which are the primary cause for developing cancer. However, polyp segmentation challenging task due diverse shape, size, color, and texture of polyps, shuttle difference between its background, as well low contrast colonoscopic images. To address these challenges, we propose feature enhancement network accurate in colonoscopy Specifically, proposed enhances semantic information using novel Semantic Feature Enhance Module (SFEM)....
Colorectal cancer (CRC) is one of the most common types with a high mortality rate. Colonoscopy preferred procedure for CRC screening and has proven to be effective in reducing mortality. Thus, reliable computer-aided polyp detection classification system can significantly increase effectiveness colonoscopy. In this paper, we create an endoscopic dataset collected from various sources annotate ground truth location results help experienced gastroenterologists. The serve as benchmark platform...
Regularized autoencoders learn the latent codes, a structure with regularization under distribution, which enables them capability to infer codes given observations and generate new samples codes. However, they are sometimes ambiguous as tend produce reconstructions that not necessarily faithful reproduction of inputs. The main reason is enforce learned code distribution match prior while true remains unknown. To improve reconstruction quality space manifold structure, this paper presents...
Polyp has long been considered as one of the major etiologies to colorectal cancer which is a fatal disease around world, thus early detection and recognition polyps plays an crucial role in clinical routines. Accurate diagnoses through endoscopes operated by physicians becomes chanllenging task not only due varying expertise physicians, but also inherent nature endoscopic inspections. To facilitate this process, computer-aid techniques that emphasize on fully-conventional image processing...
The paper proposes a Dynamic ResBlock Generative Adversarial Network (DRB-GAN) for artistic style transfer. code is modeled as the shared parameters ResBlocks connecting both encoding network and transfer network. In network, class-aware attention mechanism used to attend feature representation generating codes. multiple are designed integrate extracted CNN semantic then feed into spatial window Layer-Instance Normalization (SW-LIN) decoder, which enables high-quality synthetic images with...
Abstract Pollinators are undergoing a global decline. Although vital to pollinator conservation and ecological research, species-level identification is expensive, time consuming, requires specialized taxonomic training. However, deep learning computer vision providing ways open this methodological bottleneck through automated from images. Focusing on bumble bees, we compare four convolutional neural network classification models evaluate prediction speed, accuracy, the potential of...
Colorectal cancer is the third most common diagnosed in both men and women United States. Most colorectal cancers start as a growth on inner lining of colon or rectum, called 'polyp'. Not all polyps are cancerous, but some can develop into cancer. Early detection recognition type critical to prevent change outcomes. However, visual classification challenging due varying illumination conditions endoscopy, variant texture, appearance, overlapping morphology between polyps. More importantly,...
The maturity level of tomato is a key factor picking, which directly determines the transportation distance, storage time, and market freshness postharvest tomato. In view lack studies on classification under nature greenhouse environment, this paper proposes SE-YOLOv3-MobileNetV1 network to classify four kinds maturity. proposed model improved in terms speed accuracy: (1) Speed: Depthwise separable convolution used. (2) Accuracy: Mosaic data augmentation, K-means clustering algorithm,...
In this report, we summarize the first NTIRE challenge on light field (LF) image super-resolution (SR), which aims at super-resolving LF images under standard bicubic degradation with a magnification factor of 4. This develops new dataset called NTIRE-2023 for validation and test, provides toolbox BasicLFSR to facilitate model development. Compared single SR, major SR lies in how exploit complementary angular information from plenty views varying disparities. total, 148 participants have...
As a representative of the information revolution, Internet began in 1960s when ARPANET was born, and after decades development evolution, today it has formed global connection interaction network. The greatest value is that everyone can communicate cooperate real time across limitations space, which greatly improves efficiency group communication collaboration, then changes organization operation mode people's work, business social activities, ultimately promotes advancement human...
cloud computing (cloud computing) is a kind of distributed computing, referring to the network "cloud" will be huge data calculation and processing program into countless small programs, then, through system composed multiple servers process analyze these programs get results return user. This report explores intersection financial information processing, identifying risks challenges faced by institutions in adopting technology. It discusses need for intelligent solutions enhance efficiency...
With the rapid development of information technology, multiple time series forecasting, which is typical traffic flow has become increasingly important in big data analysis. As cornerstone intelligent transportation system, forecasting scientific research value and practical application for urban operation scheduling, quality efficiency improvement logistics industry public travel planning. Traffic prediction always an task system. Due to complex temporal spatial dependence sequence, it very...
Abstract Topology optimization (TO) method is developed and applied to 3D heat sinks design using the finite element (FEM). To go from theory practical application of optimized sink designs, impact ratio between fin height (Hc) substrate thickness (Hb) on performance was investigated, studying critical role geometric parameters in dissipation performance, focusing optimizing internal fluid channels enhance overall hydraulic thermal efficiencies under various operating conditions. Optimal...
Traffic surveillance is an important topic in computer vision and intelligent transportation systems has intensively been studied the past decades. However, most of state-of-the-art methods concentrate on daytime traffic monitoring. In this paper, we propose a nighttime system, which consists headlight detection, tracking pairing, camera calibration vehicle speed estimation. First, detected using reflection intensity map suppressed based analysis light attenuation model. Second, tracked...