A5100423076- Visual Attention and Saliency Detection
- Advanced Image and Video Retrieval Techniques
- Caching and Content Delivery
- Video Analysis and Summarization
- Human Pose and Action Recognition
- Advanced Data Storage Technologies
- Advanced Neural Network Applications
- Multimodal Machine Learning Applications
- Software Testing and Debugging Techniques
- Advanced Multi-Objective Optimization Algorithms
- Metaheuristic Optimization Algorithms Research
- Advanced Vision and Imaging
- Cloud Computing and Resource Management
- Software Reliability and Analysis Research
- IoT and Edge/Fog Computing
- Software Engineering Research
- Wireless Signal Modulation Classification
- Cellular Automata and Applications
- Image and Video Quality Assessment
- Video Coding and Compression Technologies
- Data Stream Mining Techniques
- Human Motion and Animation
- Robotic Path Planning Algorithms
- Error Correcting Code Techniques
- Distributed Sensor Networks and Detection Algorithms
Washington University in St. Louis
2024
Southwest University
2021-2024
Harbin University of Commerce
2024
Xidian University
2010-2023
Naval Aeronautical and Astronautical University
2023
Shandong Sport University
2023
Communication University of China
2020-2023
Chongqing University of Posts and Telecommunications
2022-2023
Nanjing University of Aeronautics and Astronautics
2023
Xi'an University of Technology
2018-2023
Effective convolutional features play an important role in saliency estimation but how to learn powerful for is still a challenging task. FCN-based methods directly apply multi-level without distinction, which leads sub-optimal results due the distraction from redundant details. In this paper, we propose novel attention guided network selectively integrates contextual information progressive manner. Attentive generated by our can alleviate of background thus achieve better performance. On...
Effective integration of contextual information is crucial for salient object detection. To achieve this, most existing methods based on 'skip' architecture mainly focus how to integrate hierarchical features Convolutional Neural Networks (CNNs). They simply apply concatenation or element-wise operation incorporate high-level semantic cues and low-level detailed information. However, this can degrade the quality predictions because cluttered noisy also be passed through. address problem, we...
Recent research in 4D saliency detection is limited by the deficiency of a large-scale light field dataset. To address this, we introduce new dataset to assist subsequent detection. best our knowledge, this date largest which provides 1465 all-focus images with human-labeled ground truth masks and corresponding focal stacks for every image. verify effectiveness data, first fusion framework includes two CNN streams where serve as input. The stack stream utilizes recurrent attention mechanism...
In this paper, a new meta-heuristic algorithm, called beetle swarm optimization (BSO) is proposed by enhancing the performance of through foraging principles. The 23 benchmark functions tested and compared with widely used algorithms, including particle (PSO) genetic algorithm (GA) grasshopper (GOA). Numerical experiments show that BSO outperforms its counterparts. Besides, to demonstrate practical impact two classic engineering design problems, namely, pressure vessel problem himmelblau?s...
With the evolutionary development of modern communications technology, automatic modulation classification (AMC) has played an increasing role in complex wireless communication environment. Existing AMC schemes based on deep learning use a neural network to extract features and calculate feature maps, then feed them into fully connected layers for classification. However, existing still are insufficient utilizing maps. To overcome this limitation, novel adaptive wavelet (AWN) is proposed,...
Researchers have recently achieved significant advances in deep learning techniques, which turn has substantially advanced other research disciplines, such as natural language processing, image speech recognition, and software engineering. Various techniques been successfully employed to facilitate engineering tasks, including code generation, refactoring, fault localization. Many papers also presented top conferences journals, demonstrating the applications of resolving various tasks....
Referring expressions are natural language descriptions that identify a particular object within scene and widely used in our daily conversations. In this work, we focus on segmenting the an image specified by referring expression. To end, propose end-to-end trainable comprehension network consists of visual encoders to extract feature representations from both domains. We introduce spatial-aware dynamic filters transfer knowledge text image, effectively capture spatial information object....
Background Studies have explored the relationship between social class and health for decades. However, underlying mechanism two remains not fully understood. This study aimed to explore whether self-management had a mediating role under framework of Socio-cultural Self Model. Methods 663 adults, randomly sampled from six communities in Southwest China, completed survey this study. Social was assessed using individuals’ income, education, occupation. Health through evaluation behavior,...
Automatic modulation classification (AMC) is a crucial stage in the spectrum management, signal monitoring, and control of wireless communication systems. The accurate format plays vital role subsequent decoding transmitted data. End-to-end deep learning methods have been recently applied to AMC, outperforming traditional feature engineering techniques. However, AMC still has limitations low signal-to-noise ratio (SNR) environments. To address drawback, we propose novel AMC-Net that improves...
In this paper, a new meta-heuristic algorithm, called beetle swarm optimization is proposed by enhancing the performance of through foraging principles. The 23 benchmark functions tested and compared with widely used algorithms, including particle genetic algorithm (GA) grasshopper . Numerical experiments show that outperforms its counterparts. Besides, to demonstrate practical impact two classic engineering design problems, namely, pressure vessel problem himmelblaus problem, are also...
In this paper, the path planning problem of goods transportation is formulated as a traveling salesman (TSP). A novel algorithm for warehouse robots based on two-dimensional (2D) grid model proposed to solve type TSP. Firstly, we simplified traditional pile node-based 2D model. Then, new concept called largest convex polygon (LCP) introduced illustrate shortest traverse all locations in an ideal condition. Next, remaining are classified by their relationship with LCP and designed rules...
Learning temporally consistent foreground opacity from videos, i.e., video matting, has drawn great attention due to the blossoming of conferencing. Previous approaches are built on top image matting models, which fail in maintaining temporal coherence when being adapted videos. They either utilize optical flow smooth frame-wise prediction, where performance is dependent selected model; or naively combine feature maps multiple frames, does not model well correspondence pixels adjacent...
Abstract An independent university refers to an institution of higher learning that has a bachelor's degree or education, and cooperates with social organizations other than state institutions individuals use non-state financial funds establish implement undergraduate education. While welcoming the arrival fifth-generation communication technology (5G) era, China's universities are also carrying out educational reforms adapt development. 5G created new opportunities for development students....
Saliency detection task has witnessed a booming interest for years, due to the growth of computer vision community. In this paper, we introduce new saliency model that performs active learning with kernelized subspace ranker (KSR) referred as KSR-AL. This pool-based algorithm ranks informativeness unlabeled data by considering both uncertainty sampling and information density, thereby minimizing cost labeling. The informative images are selected train KSR iteratively incrementally. is...
In this project the EEG – electroencephalogram - channel(s) will be characterized to diagnose PTSD Post-traumatic stress disorder cases. For aim, we applied boosting methods including a combination of K-mean and Support Vector Machine (SVM) models find feature weights detect We classified 32 channels 12 subjects 6 healthy controls using 6-mean classifier. The linear SVM found distinguished within each subject for cluster. It was that significant F4, F8, Pz are smaller in than subjects. This...
Recommendation bias towards objects has been found to have an impact on personalized recommendation, since present heterogeneous characteristics in some network-based recommender systems. In this article, based a biased heat conduction recommendation algorithm (BHC) which considers the heterogeneity of target objects, we propose (HHC), by further taking source into account. Tested three real datasets, Netflix, RYM and MovieLens, HHC is better both accuracy diversity than two benchmark...
Summary With the rapid development of Internet Things (IoT), fog computing has emerged as a complementary solution to address issues faced in cloud computing. However, it is challenging issue ensure both high Quality Service (QoS) and low cost for different requests when dealing with resources. In this article, we propose new approach on adaptive cost‐efficient QoS‐aware application placement called DATSS. Specifically, consists QoS state driven strategy credibility rating mechanism. The...
Recently, the problem of security distributed estimation in adversarial multitask wireless sensor networks has attracted extensive attention. For example, malicious attackers always affect signal processing and reduce network performance by destroying data information. To address this issue, using task similarity, article proposes a diffusion least-mean square algorithm based on Bayes (BDLMS) environment. The BDLMS can divide system into two subsystems: Noncooperative LMS (NCLMS) subsystem...
Differential evolution (DE) has shown remarkable performance in solving continuous optimization problems. However, its still encounters limitations when confronted with complex problems lots of local regions. To address this issue, paper proposes a dual elite groups-guided mutation strategy called “DE/current-to-duelite/1” for DE. As result, novel DE variant DEGGDE is developed. Instead only using the elites current population to direct all individuals, additionally maintains an archive...