-
1دورية أكاديمية
المؤلفون: Yuan, Weimin, Meng, Cai, Bai, Xiangzhi
المصدر: Pattern Recognition ; volume 146, page 110006 ; ISSN 0031-3203
مصطلحات موضوعية: Artificial Intelligence, Computer Vision and Pattern Recognition, Signal Processing, Software
الإتاحة: https://doi.org/10.1016/j.patcog.2023.110006Test
https://api.elsevier.com/content/article/PII:S0031320323007033?httpAccept=text/xmlTest
https://api.elsevier.com/content/article/PII:S0031320323007033?httpAccept=text/plainTest -
2دورية أكاديمية
المؤلفون: Meng, Cai, Xu, Weidong, Li, Haiyuan, Zhang, Huan, Xu, Dong
المصدر: International Journal of Advanced Robotic Systems ; volume 14, issue 3, page 172988141770738 ; ISSN 1729-8814 1729-8814
مصطلحات موضوعية: Artificial Intelligence, Computer Science Applications, Software
-
3دورية أكاديمية
المؤلفون: Meng, Cai, Li, Yanggang, Xu, Yizhou, Li, Ning, Xia, Kun
المساهمون: National Key Research and Development Program of China, National Natural Science Foundation of China
المصدر: IEEE Transactions on Medical Robotics and Bionics ; volume 4, issue 2, page 381-390 ; ISSN 2576-3202
مصطلحات موضوعية: Artificial Intelligence, Control and Optimization, Computer Science Applications, Human-Computer Interaction, Biomedical Engineering
الإتاحة: https://doi.org/10.1109/tmrb.2022.3171310Test
http://xplorestaging.ieee.org/ielx7/8253409/9777833/09765532.pdf?arnumber=9765532Test -
4
المؤلفون: Peter E Metcalfe, Yang Wang, Mikel Byrne, Ben Archibald-Heeren, Anatoly B. Rosenfeld, Meng Cai
المصدر: Physical and Engineering Sciences in Medicine. 43:1161-1170
مصطلحات موضوعية: Offset (computer science), Computer science, Biomedical Engineering, Biophysics, Iterative reconstruction, Imaging phantom, Image Processing, Computer-Assisted, Radiology, Nuclear Medicine and imaging, Computer vision, MATLAB, Instrumentation, Cone beam ct, computer.programming_language, Radiological and Ultrasound Technology, Phantoms, Imaging, business.industry, Cone-Beam Computed Tomography, Proof of concept, Beam hardening, Artificial intelligence, Artifacts, business, computer, Algorithms, Decoupling (electronics), Biotechnology
الوصف: Cone-beam computed tomography (CBCT) is an important imaging modality for image-guided radiotherapy and adaptive radiotherapy. Feldkamp-Davis-Kress (FDK) method is widely adopted in clinical CBCT reconstructions due to its fast and robust application. While iterative algorithms have been shown to outperform FDK techniques in reducing noise and imaging dose, they are unable to correct projection-domain artefacts such as beam hardening and scatter. Empirical correction techniques require a holistic approach as beam hardening and scatter coexist in the measurement data. This multi-part proof of concept study conducted in MATLAB presents a novel approach to artefact reduction for CBCT image reconstruction. Firstly, we decoupled the beam hardening and scatter contributions originating from the imaging object and the bowtie filter. Next, a model was constructed to apply pixel-wise corrections to separately account for artefacts induced by the imaging object and the bowtie filter, in order to produce mono-energetic equivalent and scatter-compensated projections. Finally, the effectiveness of the correction model was tested on an offset phantom scan as well as a clinical brain scan. A conjugate-gradient least-squares algorithm was implemented over five iterations using FDK result as the initial input. Our proposed correction model was shown to effectively reduce cupping and shading artefacts in both phantom and clinical studies. This simple yet effective correction model could be readily implemented by physicists seeking to explore the benefits of iterative reconstruction.
الوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::0e38d5bcfec27b9904cf3d1f09292cb9Test
https://doi.org/10.1007/s13246-020-00918-8Test -
5
المؤلفون: Jianxun Li, Yunnan Wang, Meng Cai
المصدر: 2021 40th Chinese Control Conference (CCC).
مصطلحات موضوعية: Computer science, business.industry, Knowledge engineering, Machine learning, computer.software_genre, Semantics, Data modeling, law.invention, Convolutional code, law, Key (cryptography), Code (cryptography), Segmentation, Artificial intelligence, business, computer, Distillation
الوصف: Deep convolutional networks have achieved great success in semantic segmentation. However, existing semantic segmentation networks are difficult to achieve an ideal balance between accuracy and speed. To tackle this dilemma, the key idea of this paper is to design a two-part knowledge distillation framework to improve the accuracy of real-time networks, including separated distillation and long-range distillation. Separated distillation treats semantic segmentation as an independent pixel classification task, aiming to align logits between a cumbersome teacher network and a lightweight student network. To extract the long-range knowledge from the scene, the similarity attention map is defined as knowledge, which is transfered from teacher to student through the long-range distillation. The framework proposed in this paper improves the accuracy of real-time networks on Cityscapes and Fisheye datasets by 2.7% and 1.7%, respectively. The distilled network is tested on NVIDIA Xavier embedded platform with speed of 15FPS. Code will be available here: https://github.com/WangYunnan/Real-Time-Semantic-SegmentationTest.
الوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_________::3938025e0f7423ac5a9fdfaf0eb96a43Test
https://doi.org/10.23919/ccc52363.2021.9550007Test -
6
المؤلفون: Yaming Liu, Zejun Ma, Meng Cai, Haitao Yao, Yao Tian
المصدر: ICASSP
مصطلحات موضوعية: Artificial neural network, business.industry, Computer science, Machine learning, computer.software_genre, Recurrent neural network, Connectionism, Keyword spotting, Artificial intelligence, Transfer of learning, business, computer, Encoder, Decoding methods, Network model
الوصف: The recurrent neural network transducer (RNN-T) model has been proved effective for keyword spotting (KWS) recently. However, compared with cross-entropy (CE) or connectionist temporal classification (CTC) based models, the additional prediction network in the RNN-T model increases the model size and computational cost. Besides, since the keyword training data usually only contain the keyword sequence, the prediction network might has over-fitting problems. In this paper, we improve the RNN-T modeling for small-footprint keyword spotting in three aspects. First, to address the over-fitting issue, we explore multi-task training where a CTC loss is added to the encoder. The CTC loss is calculated with both KWS data and ASR data, while the RNN-T loss is calculated with ASR data so that only the encoder is augmented with KWS data. Second, we use the feed-forward neural network to replace the LSTM for prediction network modeling. Thus all possible prediction network outputs could be pre-computed for decoding. Third, we further improve the model with transfer learning, where a model trained with 160 thousand hours of ASR data is used to initialize the KWS model. On a self-collected far-field wake-word testset, the proposed RNN-T system greatly improves the performance comparing with a strong "keyword-filler" baseline.
الوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_________::236dacd8f5bc9eeb62bd8ba19858763fTest
https://doi.org/10.1109/icassp39728.2021.9414339Test -
7
المؤلفون: Meng Cai
المصدر: Heliyon, Vol 7, Iss 3, Pp e06322-(2021)
Heliyonمصطلحات موضوعية: 0301 basic medicine, Exploit, Text mining, Computer science, Big data, Urban studies, Review Article, computer.software_genre, 03 medical and health sciences, 0302 clinical medicine, lcsh:Social sciences (General), lcsh:Science (General), Multidisciplinary, Land use, business.industry, Natural language processing, Urban design, Usability, Urban big data, Urban research, 030104 developmental biology, Research questions, lcsh:H1-99, Artificial intelligence, business, computer, 030217 neurology & neurosurgery, lcsh:Q1-390
الوصف: Natural language processing (NLP) has shown potential as a promising tool to exploit under-utilized urban data sources. This paper presents a systematic review of urban studies published in peer-reviewed journals and conference proceedings that adopted NLP. The review suggests that the application of NLP in studying cities is still in its infancy. Current applications fell into five areas: urban governance and management, public health, land use and functional zones, mobility, and urban design. NLP demonstrates the advantages of improving the usability of urban big data sources, expanding study scales, and reducing research costs. On the other hand, to take advantage of NLP, urban researchers face challenges of raising good research questions, overcoming data incompleteness, inaccessibility, and non-representativeness, immature NLP techniques, and computational skill requirements. This review is among the first efforts intended to provide an overview of existing applications and challenges for advancing urban research through the adoption of NLP.
Natural language processing; Urban research; Urban big data; Text miningالوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::8f5eca6b4382377ae6bdf05507aa37bfTest
http://www.sciencedirect.com/science/article/pii/S2405844021004278Test -
8دورية أكاديمية
المؤلفون: Eva Kassens-Noor (10037534), Mark Wilson (734683), Zeenat Kotval-Karamchandani (10037537), Meng Cai (313811), Travis Decaminada (10037540)
مصطلحات موضوعية: Urban and Regional Planning not elsewhere classified, artificial intelligence, domotics, autonomous vehicles, perception, willingness, inteligencia artificial, domótica, vehículos autónomos, percepción, disposición, 人工智能, 家庭自动化, 自动驾驶汽车, 感知力, 意愿
الوصف: Supplemental material, sj-docx-1-jpe-10.1177_0739456X20984529 for Living with Autonomy: Public Perceptions of an AI-Mediated Future by Eva Kassens-Noor, Mark Wilson, Zeenat Kotval-Karamchandani, Meng Cai and Travis Decaminada in Journal of Planning Education and Research
-
9
المؤلفون: Ming Yao, Yu-meng Cai, Chen Shi'an, Jian Wang
المصدر: International Journal of Control, Automation and Systems. 16:2203-2213
مصطلحات موضوعية: Normalization (statistics), 0209 industrial biotechnology, Differential equation, business.industry, Computer science, Robotics, 02 engineering and technology, Mechatronics, Active suspension, Linear-quadratic-Gaussian control, 01 natural sciences, Computer Science Applications, Axle, 020901 industrial engineering & automation, Control and Systems Engineering, Control theory, 0103 physical sciences, Artificial intelligence, Minimum phase, business, 010301 acoustics
الوصف: To improve vehicle roll safety during steering operation at high running speed, a new design approach for Linear Quadratic Gaussian (LQG) controller of active suspension system is proposed. Key steps of the new approach are as follows: 1) The front axle steered angle is written into a differential equation in accord with the minimum phase system and combines with the original system into the augmented system equation; 2) positive infinitesimals respectively including controls are added to the index; Thirdly, weights of the evaluating indicators of the LQG controller are determined by using analytic hierarchy process (AHP) and normalization methods based on vehicle motion statistics under the double-lane change maneuver as the typical steering maneuver. Performance comparisons are implemented between the active suspension system and the passive one under the double-lane change, slalom and fish-hook maneuvers. Results verify that the active suspension system with the proposed controller can achieve better vehicle roll safety and has a good adaptability under different steering maneuvers.
الوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_________::f6f8b14b468eee39402729a66e1c0df4Test
https://doi.org/10.1007/s12555-017-0159-2Test -
10
المؤلفون: Xinlin He, Jianxun Li, Meng Cai
المصدر: 2020 Chinese Control And Decision Conference (CCDC).
مصطلحات موضوعية: Clustering high-dimensional data, 0209 industrial biotechnology, Training set, business.industry, Computer science, Posterior probability, Pattern recognition, 02 engineering and technology, Latent variable, Synthetic data, 020901 industrial engineering & automation, 0202 electrical engineering, electronic engineering, information engineering, Probability distribution, Oversampling, 020201 artificial intelligence & image processing, Artificial intelligence, business, Classifier (UML), Encoder
الوصف: The imbalance classification problem arises when certain class is underrepresented in comparison with other classes, leading to a classifier partial to the majority classes. Existing interpolation-based oversampling methods for handling this problem characteristically do not make full use of the probability distribution of data. To overcome this weakness, this study proposes latent posterior based generative adversarial network oversampling approach(LPGOS), which uses a variational encoder to obtain the posterior distribution of latent variables. In addition, giving the high correlation between generated synthetic data and original data, we introduce a transfer learning approach with weight scaling factor namely TrWSBoost in which the generated minority class samples are treated as source domain data. Visual results prove that the proposed approach LPGOS is capable of approximate high dimensional data distribution and outperform other existing oversampling techniques. The performance of binary classifiers verify the effectiveness of proposed approaches.
الوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_________::2f8d32fce6cf2548e12807747b9ac804Test
https://doi.org/10.1109/ccdc49329.2020.9164226Test