A5064702700- Video Surveillance and Tracking Methods
- Generative Adversarial Networks and Image Synthesis
- Human Pose and Action Recognition
- Image Enhancement Techniques
- Advanced Image Processing Techniques
- Advanced Vision and Imaging
- Face recognition and analysis
- Face and Expression Recognition
- Medical Image Segmentation Techniques
- Brain Tumor Detection and Classification
- Visual Attention and Saliency Detection
- Advanced Image and Video Retrieval Techniques
- Image and Signal Denoising Methods
- Fire Detection and Safety Systems
- Anomaly Detection Techniques and Applications
- Advanced Neural Network Applications
- Facial Nerve Paralysis Treatment and Research
- Mobile Crowdsensing and Crowdsourcing
- Expert finding and Q&A systems
- Impact of Light on Environment and Health
- Digital Media Forensic Detection
- Advanced Image Fusion Techniques
- Recommender Systems and Techniques
- Topic Modeling
- Infrared Target Detection Methodologies
Huazhong Agricultural University
2015-2024
Huazhong University of Science and Technology
2009-2014
Abstract Cotton is an important economic crop, and many loci for traits have been identified, but it remains challenging time-consuming to identify candidate or causal genes/variants clarify their roles in phenotype formation regulation. Here, we first collected integrated the multi-omics datasets including 25 genomes, transcriptomes 76 tissue samples, epigenome data of five species metabolome 768 metabolites from four tissues, genetic variation, trait transcriptome 4180 cotton accessions....
Dictionary learning for sparse representation has been increasingly applied to object tracking, however, the existing methods only utilize one modality of learn a single dictionary. In this paper, we propose robust tracking method based on multitask joint dictionary learning. Through extracting different features target, multiple linear representations are obtained. Each can be learned by corresponding Instead separately dictionaries, adopt approach representations, which provide additional...
Deep learning has powered many face related tasks and shown state-of-the-art performance. However, existing deep models are often trained separately for different problems, which results in heavy computational burden. To address this problem, we propose a novel multi-task network with fully convolutional architecture-Hierarchical Multi-task Network (HMT-Net), that simultaneously recognizes person's gender, race facial attractiveness from given portrait image. Aiming to improve the robustness...
Feature extraction plays a significant part in computer vision tasks. In this paper, we propose method which transfers rich deep features from pretrained model on face verification task and feeds the into Bayesian ridge regression algorithm for facial beauty prediction. We leverage neural networks that extracts more abstract stacked layers. Through simple but effective feature fusion strategy, our achieves improved or comparable performance SCUT-FBP dataset ECCV HotOrNot dataset. Our...
Moving object detection is a fundamental step in video surveillance system. To eliminate the influence of illumination change and shadow associated with moving objects, we proposed local intensity ratio model (LIRM) which robust to change. Based on analysis model, discussed distribution ratio. And objects are segmented without using normalized via Gaussian mixture (GMM). Then erosion used get contours erase scatter patches noises. After that, enhanced by new contour enhancement method,...
Correlation filter has drawn increasing interest in visual tracking due to its high efficiency, however, it is sensitive partial occlusion, which may result failure. To address this problem, we propose a novel local-global correlation (LGCF) for object tracking. Our LGCF model utilizes both local-based and global-based strategies, effectively combines these two strategies by exploiting the relationship of circular shifts among local parts global target their motion models preserve structure...
Real-world actions occur often in crowded, dynamic environments. This poses a difficult challenge for current approaches to video event detection because crowd scenes are always extremely cluttered. In this paper, we design content analysis method fighting recognition scene. Our begins with four MPEG-7 descriptors: kinetic energy, motion directions histogram, spatial distribution parameter and localization of two adjacent frames. Then the support vector machines (SVMs) is introduced train...
The Bag of Visual Words (BoW) model is one the most popular and effective image classification frameworks in recent literature. optimal formation a visual vocabulary remains unclear, size also affects performance classification. Empirically, larger leads to higher accuracy. However, needs more memory intensive computational resources. In this paper, we propose multiresolution feature coding (MFC) framework via aggregating codings obtained from set small vocabularies with different sizes,...
Loss function is crucial for model training and feature representation learning, conventional models usually regard facial attractiveness recognition task as a regression problem, adopt MSE loss or Huber variant supervision to train deep convolutional neural network (CNN) predict score. Little work has been done systematically compare the performance of diverse functions. In this paper, we firstly analyze under Then novel named ComboLoss proposed guide SEResNeXt50 network. The method...
<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2"><mml:mrow><mml:mi>k</mml:mi></mml:mrow></mml:math>-means algorithm is a widely used clustering in data mining and machine learning community. However, the initial guess of cluster centers affects result seriously, which means that improper initialization cannot lead to desirous result. How choose suitable an important research issue for<mml:math id="M3"><mml:mrow><mml:mi>k</mml:mi></mml:mrow></mml:math>-means algorithm. In this...
We propose a novel part-based tracking algorithm using online weighted P-N learning. An learning method is implemented via considering the weight of samples during classification, which improves performance classifier. apply to track target model instead whole target. In doing so, object segmented into fragments and parts them are selected as local feature blocks (LFBs). Then, employed train classifier for each block (LFB). Each LFB tracked through corresponding classifier, respectively....
Abnormal behavior detection is an important issue in video surveillance. This paper presents approach for abnormal based on spatial-temporal features. First, the proposed method extracts moving objects from sequence. Then, it tracks to detect their overlapping. Finally, a clutter-model built up changes of feature behavior. Experimental results show effectiveness approach.
Integrating visible and infrared images into one high-quality image, also known as image fusion, is a challenging yet critical task for many downstream vision tasks. Most existing works utilize pretrained deep neural networks or design sophisticated frameworks with strong priors this task, which may be unsuitable lack flexibility. This paper presents SimpleFusion, simple effective framework fusion. Our follows the decompose-and-fusion paradigm, where are decomposed reflectance illumination...
Reconstruction-based methods have significantly advanced modern unsupervised anomaly detection. However, the strong capacity of neural networks often violates underlying assumptions by reconstructing abnormal samples well. To alleviate this issue, we present a simple yet effective reconstruction framework named Attention-Guided Pertuation Network (AGPNet), which learns to add perturbation noise with an attention mask, for accurate Specifically, it consists two branches, \ie, plain branch and...
Person re-identification (ReID) has recently benefited from large pretrained vision-language models such as Contrastive Language-Image Pre-Training (CLIP). However, the absence of concrete descriptions necessitates use implicit text embeddings, which demand complicated and inefficient training strategies. To address this issue, we first propose one straightforward solution by leveraging existing image captioning to generate pseudo captions for person images, thereby boost with vision...