A5101829754- Natural Language Processing Techniques
- Topic Modeling
- Web Data Mining and Analysis
- Text and Document Classification Technologies
- Semantic Web and Ontologies
- Multimodal Machine Learning Applications
- Advanced Text Analysis Techniques
- Video Surveillance and Tracking Methods
- Spam and Phishing Detection
- Human Mobility and Location-Based Analysis
- Traffic Prediction and Management Techniques
- Complex Network Analysis Techniques
- Thyroid Cancer Diagnosis and Treatment
- Advanced Image and Video Retrieval Techniques
- Hearing Impairment and Communication
- Hand Gesture Recognition Systems
- Advanced Vision and Imaging
- Anomaly Detection Techniques and Applications
- Sentiment Analysis and Opinion Mining
- Data Quality and Management
- Advanced Computational Techniques and Applications
- Advanced Neural Network Applications
- Visual Attention and Saliency Detection
- Scientific Computing and Data Management
- Stock Market Forecasting Methods
Southwestern University of Finance and Economics
2022-2025
University of Electronic Science and Technology of China
2018-2024
Wuhan University of Science and Technology
2013-2024
Sichuan Academy of Medical Sciences & Sichuan Provincial People's Hospital
2024
South China Normal University
2022
Beijing Aerospace Flight Control Center
2022
China Railway Construction Machinery Research & Design Institute
2022
Hainan Medical College Hospital
2020
Chengdu University of Technology
2011-2019
Nanjing General Hospital of Nanjing Military Command
2019
Image semantic segmentation has always been a research hotspot in the field of robots. Its purpose is to assign different category labels objects by segmenting objects. However, practical applications, addition knowing information objects, robots also need know position complete more complex visual tasks. Aiming at indoor environment, this study designs an image network framework joint target detection. Using parallel operation adding branches detection network, it innovatively implements...
In this paper, we introduce a web-scale general visual search system deployed in Microsoft Bing. The accommodates tens of billions images the index, with thousands features for each image, and can respond less than 200 ms. order to overcome challenges relevance, latency, scalability such large scale data, employ cascaded learning-to-rank framework based on various latest deep learning features, deploy distributed heterogeneous computing platform. Quantitative qualitative experiments show...
This paper reviews the first AIM challenge on bokeh effect synthesis with focus proposed solutions and results. The participating teams were solving a real-world image-to-image mapping problem, where goal was to map standard narrow-aperture photos same captured shallow depth-of-field by Canon 70D DSLR camera. In this task, participants had restore based only one single frame without any additional data from other cameras or sensors. target metric used in combined fidelity scores (PSNR SSIM)...
Abstract Security issues in cloud computing have become a hot topic academia and industry, CP-ABE is an effective solution for managing protecting data. When data shared computing, they usually multiple access structures that hierarchical relationships. However, existing algorithms do not consider such relationships just require owners to generate ciphertexts meet the requirement, which would incur substantial computation overheads. To achieve fine-grained control of files effectively, first...
Summary As the development of deep learning and continuous improvement computing power, as well needs social production, target detection has become a research hotspot in recent years. However, algorithm problem that it is more sensitive to large targets does not consider feature‐feature interrelationship, which leads high false or missed rate small targets. An method (C‐SSD) based on improved SSD proposed, replaces backbone network VGG‐16 with dense convolution (C‐DenseNet) achieves further...
Recent progress in text classification has been focused on high-resource languages such as English and Chinese. For low-resource languages, amongst them most African the lack of well-annotated data effective preprocessing, is hindering transfer successful methods. In this paper, we introduce two news datasets (KINNEWS IRNEWS) for multi-class articles Kinyarwanda Kirundi, languages. The are mutually intelligible, but while studied Natural Language Processing (NLP) to some extent, work...
Text categorization is the process of assigning documents to a set previously fixed categories. It widely used in many data-oriented management applications. Many popular algorithms for text have been proposed, such as Naive Bayes, k-Nearest Neighbor (k-NN), Support Vector Machine (SVM). However, those classification approaches do not perform well every case, example, SVM can identify categories correctly when texts are cross zones multi-categories, k-NN cannot effectively solve problem...
Social media offer abundant information for studying people's behaviors, emotions and opinions during the evolution of various rare events such as natural disasters. It is useful to analyze correlation between social human-affected events. This study uses Hurricane Sandy 2012 related Twitter text data conduct extraction classification. Considering that original contains different topics, we need find Sandy. A fuzzy logic-based approach introduced solve problem Inputs used in proposed model...
Due to homonyms, abbreviations, etc., name ambiguity is widely available in Web and e-document. For example, when integrating heterogeneous literature databases, because there are different specifications, authors may be thought of as the same author, vice versa. Therefore, makes data robust even dirty lowers precision information retrieval. In this paper, we present an approach, named semantic association based disambiguation method (SAND), solve person ambiguity. The basic idea SAND...
In this work, we study the problem of recognizing identification (ID) information from unconstrained real-world images ID card, which has extensively applied in practical scenarios. Nonetheless, manual ways processing task are impractical due to unaffordable cost labor and time consumption as well unreliable quality labeling. paper, propose an intelligent framework for automatically cards. Specifically, first conduct marginal detection using a multi-operator algorithm then localize region...
In the natural language processing family, learning representations is a pioneering study, especially in sequence-to-sequence tasks where outputs are generated, totally relying on of source sequence. Generally, classic methods infer that each word occurring sequence, having more or less influence target should all be considered when generating outputs. As summarization task requires output sequence to only retain essence, full consideration may not work well it, which calls for suitable with...