A5100414877- Handwritten Text Recognition Techniques
- Advanced Image and Video Retrieval Techniques
- Immune Response and Inflammation
- Image Retrieval and Classification Techniques
- T-cell and B-cell Immunology
- Image Processing and 3D Reconstruction
- Natural Language Processing Techniques
- Immune Cell Function and Interaction
- Generative Adversarial Networks and Image Synthesis
- Immunotherapy and Immune Responses
- NF-κB Signaling Pathways
- Chaos-based Image/Signal Encryption
- Multimodal Machine Learning Applications
- Chemical Synthesis and Characterization
- Animal Virus Infections Studies
- Zeolite Catalysis and Synthesis
- T-cell and Retrovirus Studies
- Toxin Mechanisms and Immunotoxins
- Cellular Automata and Applications
- Advanced Neural Network Applications
- Advanced oxidation water treatment
- Speech Recognition and Synthesis
- Virus-based gene therapy research
- Hand Gesture Recognition Systems
- EEG and Brain-Computer Interfaces
Nanchang University
2024
Gansu Institute of Political Science and Law
2024
Xihua University
2022-2024
Sichuan Agricultural University
2024
South China University of Technology
2019-2023
Nanjing University of Posts and Telecommunications
2023
University of Science and Technology of China
2022
University of Virginia
2021
Nanjing Normal University
2008-2019
Tianjin University of Technology
2019
Existing scene text spotting (i.e., end-to-end detection and recognition) methods rely on costly bounding box annotations (e.g., text-line, word-level, or character-level boxes). For the first time, we demonstrate that training models can be achieved with an extremely low-cost annotation of a single-point for each instance. We propose method tackles as sequence prediction task. Given image input, formulate desired recognition results discrete tokens use auto-regressive Transformer to predict...
End-to-end scene text spotting has made significant progress due to its intrinsic synergy between detection and recognition. Previous methods commonly regard manual annotations such as horizontal rectangles, rotated quadrangles, polygons a prerequisite, which are much more expensive than using single-point. Our new framework, SPTS v2, allows us train high-performing text-spotting models single-point annotation. v2 reserves the advantage of auto-regressive Transformer with an Instance...
In the face of global challenges related to Environmental, Social, and Governance (ESG) issues pressure meet Sustainable Development Goals (SDGs), green finance corporate ESG practices have become dual-driving mechanisms for promoting sustainable development. Green directs capital toward environmental protection, energy conservation, climate change mitigation, fostering development low-carbon economies technologies. Corporate practices, focusing on social responsibility, projects,...
To achieve high detectivity in all-polymer photodetectors (all-PPDs), a thick-film photoactive layer is favored because it can effectively suppress the dark current density. However, if of film too thick, leads to reduced responsivity owing increased recombination loss. We developed high-performance all-PPDs by using narrowband-gap p-type polymer NT40 and an n-type poly{[ N, N'-bis(2-octyldodecyl)-naphthalene-1,4,5,8-bis(dicarboximide)-2,6-diyl]- alt-5,5'-(2,2'bithiophene)} as layer. The...
Visual Information Extraction (VIE) has attracted considerable attention recently owing to its various advanced applications such as document understanding, automatic marking and intelligent education. Most existing works decoupled this problem into several independent sub-tasks of text spotting (text detection recognition) information extraction, which completely ignored the high correlation among them during optimization. In paper, we propose a robust System (VIES) towards real-world...
Patchoulol, a plant-derived sesquiterpene compound, is widely used in perfumes, cosmetics, and pharmaceuticals. Microbial production provides promising alternative approach for the efficient sustainable of patchoulol. However, there are no systematic engineering studies on Komagataella phaffii aimed at achieving high-yield patchoulol production. Herein, by fusion overexpression FPP synthase (ERG20LPTS), increasing precursor supply, adjusting copy number ERG20LPTS PTS, combined with adding...
In recent years, end-to-end scene text spotting approaches are evolving to the Transformer-based framework. While previous studies have shown crucial importance of intrinsic synergy between detection and recognition, advances in methods usually adopt an implicit strategy with shared query, which can not fully realize potential these two interactive tasks. this paper, we argue that explicit considering distinct characteristics recognition significantly improve performance spotting. To end,...
Motor imagery is one of the classical paradigms which have been used in brain-computer interface and motor function recovery. Finger movement-based execution a complex biomechanical architecture crucial task for establishing most complicated natural activities daily life. Some patients may suffer from alternating hemiplegia after brain stroke lose their ability execution. Fortunately, might be preserved independently worked as backdoor The efficacy achieving significant recovery cortex still...
Document layout analysis is a crucial prerequisite for document understanding, including retrieval and conversion. Most public datasets currently contain only PDF documents lack realistic documents. Models trained on these may not generalize well to real-world scenarios. Therefore, this paper introduces large diverse dataset called M <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sup> Doc. The M6 designation represents six properties: (1)...
Camera-captured document images usually suffer from perspective and geometric deformations. It is of great value to rectify them when considering poor visual aesthetics the deteriorated performance OCR systems. Recent learning-based methods intensively focus on accurately cropped image. However, this might not be sufficient for overcoming practical challenges, including either with large marginal regions or without margins. Due impracticality, users struggle crop documents precisely they...
Accurate object segmentation is crucial for high-quality scene understanding in the 3D vision domain. However, based on Gaussian Splatting (3DGS) struggles with accurately delineating boundaries, as primitives often span across edges due to their inherent volume and lack of semantic guidance during training. In order tackle these challenges, we introduce Clear Object Boundaries 3DGS Segmentation (COB-GS), which aims improve accuracy by clearly blurry boundaries interwoven within scene....
Iron is the most suitable candidate for activating peroxymonosulfate (PMS) organic degradation. To realize effective separation and recycling of iron catalysts, it urgent to explore a low-cost excellent supporter. In this study, zeolite catalyst (CAN) with both supporter functions was synthesized in one step using lithium slag (LS, containing iron), used degrade tetracycline (TC) wastewater. The results showed that electron transfer sulfur CAN promotes decomposition PMS. CAN/PMS system...
In view of the fact that current attention-recognition studies are mostly single-level-based, this paper proposes a multi-level method based on feature selection. Four experimental scenarios designed to induce high, medium, low, and non-externally directed attention states. A total 10 features extracted from electroencephalogram (EEG) channels, respectively, including time-domain measurements, sample entropy, frequency band energy ratios. Based all features, an 88.7% recognition accuracy is...