A5100389400- Handwritten Text Recognition Techniques
- Image Processing and 3D Reconstruction
- Quantum optics and atomic interactions
- Natural Language Processing Techniques
- Quantum Information and Cryptography
- Advanced Neural Network Applications
- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Data Mining Algorithms and Applications
- Technology-Enhanced Education Studies
- Quantum Mechanics and Applications
- Technology Adoption and User Behaviour
- Blockchain Technology Applications and Security
- Domain Adaptation and Few-Shot Learning
- Semantic Web and Ontologies
- Education and Learning Interventions
- Advanced Computational Techniques and Applications
- Age of Information Optimization
- Gender and Technology in Education
- Vehicle License Plate Recognition
- Video Analysis and Summarization
- Menstrual Health and Disorders
- Protein Interaction Studies and Fluorescence Analysis
- Digital literacy in education
- Digital Media Forensic Detection
Chiayi Chang Gung Memorial Hospital
2023-2024
Tsinghua Sichuan Energy Internet Research Institute
2024
Tsinghua University
2024
Beijing Academy of Artificial Intelligence
2023
Xi'an Jiaotong University
2020-2022
Guangzhou Vocational College of Science and Technology
2020
South China University of Technology
2017-2019
Texas A&M University
2017
Southeast University
2016
Zhejiang Ocean University
2016
Curved text detection is a difficult problem that has not been addressed sufficiently. To highlight the difficulties in reading curved real environment, we constructed dataset called CTW1500, which includes over 10,000 annotations 1500 images, and used it to formulate polygon-based detector can detect without using an empirical combination. With seamless integration of recurrent transverse longitudinal offset connection, our method explores context information instead predicting points...
Scene text in the wild is commonly presented with high variant characteristics. Using quadrilateral bounding box to localize instance nearly indispensable for detection methods. However, recent researches reveal that introducing scene will bring a label confusion issue which easily overlooked, and this may significantly undermine performance. To address issue, paper, we propose novel method called Sequential-free Box Discretization (SBD) by discretizing into key edges (KE) can further derive...
A new method is proposed for removing text from natural images. The challenge to first accurately localize on the stroke-level and then replace it with a visually plausible background. Unlike previous methods that require image patches erase scene text, our method, namely ensconce network (EnsNet), can operate end-to-end single without any prior knowledge. overall structure an trainable FCN-ResNet-18 conditional generative adversarial (cGAN). feature of former enhanced by novel lateral...
Professional association football is a game of talent. The success professional club hinges largely on its ability assembling the best team. Building dataset player transfer records among more than 400 clubs in 24 world-wide top class leagues from 2011 to 2015, this study aims relate club's activities market network perspective. We confirm that modern indeed money game, which larger investment spent acquisition talented players generally yields better team performance. However, further...
With the passing of history, precious cultural heritage was left behind to tell ancient stories, especially those in form written documents. In this paper, a weakly supervised segmentation system with recognition-guided information on attention area, is proposed for high-precision historical document under strict intersection-over-union (IoU) requirements. We formulate character problem from Bayesian decision theory perspective and propose boundary box (BBS), BBS (Rg-BBS), (Rg-ABBS),...
We introduce point affiliation into feature upsampling, a notion that describes the of each upsampled to semantic cluster formed by local decoder points with similarity. By rethinking affiliation, we present generic formulation for generating upsampling kernels. The kernels encourage not only smoothness but also boundary sharpness in maps. Such properties are particularly useful some dense prediction tasks such as segmentation. key idea our is generate similarity-aware comparing similarity...
The Ten-eleven translocation (TET) family of 5-methylcytosine (5mC) dioxygenases catalyze the conversion 5mC into 5-hydroxymethylcytosine (5hmC) and further oxidized species to promote active DNA demethylation. Here we engineered a split-TET2 enzyme enable temporal control oxidation subsequent remodeling epigenetic states in mammalian cells. We demonstrate use this chemically inducible system dissect correlation between hydroxymethylation chromatin accessibility genome. This...
Multipartite entanglement plays an important role in quantum information processing and metrology. Here, the dressing-energy-level-cascaded four-wave mixing (FWM) processes are proposed to generate all-optical controlled multipartite within a single device. The characteristics of produced states light characterized by applying Duan criterion positivity under partial transposition criterion. Moreover, using internal dressing field modulate atomic coherence, multiple coherent channels FWM...
The majority of kidney recipients are a subset chronic disease. Our previous study demonstrated that the combination Lactobacillus plantarum and paracasei (Lm) had highest clearance ability uremic toxins improved function in mouse model. This aimed to evaluate Lm improving graft function, effects on immunosuppressants, safety transplant recipients.We retrospectively reviewed 24 patients. Twelve them take regularly; we compared creatinine measurements estimated glomerular filtration rate 3...
Abstract Einstein–Podolsky–Rosen (EPR) steering for multipartite systems plays an important role in several quantum technologies and applications, e.g., communication metrology. In this study, energy‐level cascaded four‐wave mixing (EC‐FWM) process modulated by the dressing effect is proposed to generate three‐ four‐mode EPR atomic medium of rubidium. The simultaneously FWM ordinally are compared three‐mode steering, monogamy relations Gaussian steerability also verified. Furthermore,...
The earliest extant Chinese characters originate from oracle bone inscriptions, which are closely related to other East Asian languages. These inscriptions hold immense value for anthropology and archaeology. However, deciphering script remains a formidable challenge, with only approximately 1,600 of the over 4,500 elucidated date. Further scholarly investigation is required comprehensively understand this ancient writing system. Artificial Intelligence technology promising avenue...
In recent years, large-scale models have demonstrated state-of-the-art performance across various domains. However, training such requires techniques to address the problem of limited computing power and memory on devices as GPUs. Some commonly used include pipeline parallelism, tensor activation checkpointing. While existing works focused finding efficient distributed execution plans (Zheng et al. 2022) checkpoint scheduling (Herrmann 2019, Beaumont 2021}, there has been no method proposed...
Utilizing OnLine Short Text (OLST) in social networking tools such as microblogs, instant messaging platforms, and short message service via smart phones has become a routine daily life. OLST is appealing for personal covert communication because it can hide information very carrier text, this concealment hard to detect due the diversity of normal traffic. However, designing appropriate schemes confronts several challenges: they need be provably secure, their performance needs maintain high...
This article theorizes the functional relationship between human components (i.e., scholars) and non-human structural configurations) of academic domains. It is organized around following question: in what ways have scholars formed been by configurations their domain? The uses as a case study domain education technology to examine this question. Its authorship approach innovative, with worldwide collection academics (99 authors) collaborating address proposed question based on reflections...
This article employed modeling analysis of flower production process with Petri Net, which based on the logistics system enterprises, using Witness simulation software to study at same time. The found some existing problems through simulation, and put forward several suggestions about improvement, therefore it improved efficiency brought economic benefits. Meanwhile presents feasible schemes improvement other enterprises.
Vibration fault diagnosis of hydroelectric unit was investigated using method least square support vector machine (LS-SVM) and Dempster-Shafer theory (D-S Theory).Spectrum amplitude characteristic acted as eigenvector learning samples to train the constructed LS-SVM regression classifier for realizing mapping relationship between characteristic.Information fusion realized after completing local diagnosis, then achieved.Experiments show that has a rapidly diagnostic process generalization...
With the widespread application of mobile internet 5G and blockchain technology, research big data analysis mining has been paid more attention by enterprises, universities institutions. The use MapReduce technology to analyze explore financial transaction such as banking, insurance, securities, funds, futures foreign exchange strong requirements academic significance. This paper first analyzes characteristics transaction, then introduces related work transaction. Moreover, through...
Multimode quantum squeezing plays an essential role in the fields of metrology and information. In this paper, we first construct a three- four-mode energy-level cascaded four-wave mixing system single <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mi/> <mml:mn>85</mml:mn> </mml:mrow> </mml:msup> <mml:mi mathvariant="normal">R</mml:mi> mathvariant="normal">b</mml:mi> </mml:math> vapor, then analyze properties...
We present a depth fingerprint map to obtain three-dimensional (3D) spatial information from objects in large deterministic background by active range-gated imaging, for night remote surveillance. This method first gives the of region interest gate viewing form contour bands range. The is then embedded laser surveillance system. Finally, 3D such as target scale and location can be estimated segmenting matching them with map. measurement performed computer processing. Therefore, has no...