A5100698166- Image and Signal Denoising Methods
- Advanced Image Fusion Techniques
- Medical Image Segmentation Techniques
- Video Surveillance and Tracking Methods
- Remote-Sensing Image Classification
- Advanced Image Processing Techniques
- Image Retrieval and Classification Techniques
- Image Processing Techniques and Applications
- Remote Sensing and Land Use
- Image Enhancement Techniques
- Face and Expression Recognition
- Advanced Image and Video Retrieval Techniques
- Advanced Vision and Imaging
- Digital Media Forensic Detection
- Sparse and Compressive Sensing Techniques
- Advanced Neural Network Applications
- Bayesian Methods and Mixture Models
- Generative Adversarial Networks and Image Synthesis
- Domain Adaptation and Few-Shot Learning
- Image and Object Detection Techniques
- Image and Video Quality Assessment
- Biometric Identification and Security
- Impact of Light on Environment and Health
- Advanced Numerical Analysis Techniques
- Machine Learning and ELM
Xi'an University of Architecture and Technology
2025
Nanjing University of Information Science and Technology
2015-2024
Qinghai Normal University
2023-2024
South China Normal University
2012-2023
University of Chinese Academy of Sciences
2022
Institute of Theoretical Physics
2022
Chinese Academy of Sciences
2022
Texas Instruments (United States)
2020
Nanjing University of Science and Technology
2010-2011
Fuzzy c-means (FCM) has been considered as an effective algorithm for image segmentation. However, it still suffers from two problems: one is insufficient robustness to noise, and the other Euclidean distance in FCM, which sensitive t
Multiview learning (MVL), which enhances the learners' performance by coordinating complementarity and consistency among different views, has attracted much attention. The multiview generalized eigenvalue proximal support vector machine (MvGSVM) is a recently proposed effective binary classification method, introduces concept of MVL into classical (GEPSVM). However, this approach cannot guarantee good robustness yet. In article, we develop robust double-sided twin SVM (MvRDTSVM) with...
Kernel methods, e.g., composite kernels (CKs) and spatial-spectral (SSKs), have been demonstrated to be an effective way exploit the information nonlinearly for improving classification performance of hyperspectral image (HSI). However, these methods are always conducted with square-shaped window or superpixel techniques. Both techniques likely misclassify pixels that lie at boundaries class, thus a small target is smoothed away. To alleviate problems, in this paper, we propose novel...
Recently, the hidden Markov model (HMM) with student's t-mixture (SMM), called t-HMM (SHMM) for short, has received much attention in unsupervised learning of sequential data. However, current existing SHMMs fail to take into consideration relevant features embedded local subspaces, thus influencing their performances clustering. To address problem, a novel SHMM is proposed by combining measure localized feature saliency (LFS) SMM and utilizing two t-distributions as subcomponents...
Sequential data, such as video frames and event have been widely applied in the realworld. As a special kind of sequential hyperspectral images (HSIs) can be regarded sequence 2-D spectral dimension, which effectively utilized for distinguishing different landcovers according to sequences. This paper presents novel noise reduction method based on superpixel-based subspace low rank representation imagery. First, under framework linear mixture model, original cube is assumed domain, could...
Adverse weather conditions such as haze and snowfall can degrade the quality of captured images affect performance drone detection. Therefore, it is challenging to locate identify targets in adverse scenarios. In this paper, a novel model called Object Detection Foggy Condition with YOLO (ODFC-YOLO) proposed, which performs image dehazing object detection jointly by multi-task learning approach. Our consists subnet subnet, be trained end-to-end optimize both tasks. Specifically, we propose...
Dictionary learning has produced state-of-the-art results in various classification tasks. However, if the training data have a different distribution than testing data, learned sparse representation might not be optimal. Recently, several domain-adaptive dictionary (DADL) methods and kernels been proposed achieved impressive performance. performance of these single kernel-based heavily depends on choice kernel, question how to combine multiple kernel (MKL) with DADL framework well studied....
During the process of signal sampling and digital imaging, hyperspectral images (HSI) inevitably suffer from contamination mixed noises. The fidelity efficiency subsequent applications are considerably reduced along with this degradation. Recently, as a formidable implement for image processing, low-rank regularization has been widely extended to restoration HSI. Meanwhile, further exploration non-local self-similarity proven useful in exploiting spatial redundancy Better preservation...
The emergence of adversarial examples has had a significant impact on the development and application deep learning. In this paper, novel convolutional neural network model, stochastic multifilter statistical (SmsNet), is proposed for detection examples. A feature layer constructed to collect data map output from each in SmsNet by combining manual features with network. entire model an end-to-end so not independent network, its directly transmitted fully connected short-cut connection called...
We describe a general procedure to construct the independent and complete operator bases for generic Lorentz invariant effective field theories, given any kind of gauge symmetry content, up mass dimension. By considering as contact on-shell amplitude, so-called amplitude correspondence, we provide unified construction flavor structures by Young Tableau tensor. Several are constructed emphasize different aspects: independence (y-basis m-basis), repeated fields with flavors (p-basis),...