期刊文献+
共找到934篇文章
< 1 2 47 >
每页显示 20 50 100
A method for robust TV logo detection 预览
1
作者 潘达 Shi Ping +2 位作者 Ying Zefeng Hou Ming Han Mingliang 《高技术通讯:英文版》 CAS 2019年第2期144-152,共9页
A robust TV logo detection method based on the modified single shot multibox detector (SSD) is presented. Unlike most other existing methods which can only detect the TV logo from video frames, the proposed method can... A robust TV logo detection method based on the modified single shot multibox detector (SSD) is presented. Unlike most other existing methods which can only detect the TV logo from video frames, the proposed method can also detect the TV logo from photo pictures taken by smartphones or other smart terminals. Firstly, using a simple and effective way of collecting and labelling TV logo, a large-scale TV logo dataset used to train the detection model is built. Then, parameters and loss function of SSD are modified to make it more suitable for the task of TV logo detection. Moreover, a soft-NMS algorithm is introduced to remove the redundant overlapping boxes and obtain the final output box. And also an approach for hard example mining is designed to improve the detection accuracy. Finally, extensive comparison experiments are carried out which take into consideration different image resolutions, logo positions and environmental factors existing in real-world applications. Experimental results demonstrate that the proposed method achieve superior performances in robustness compared to other state-of-the-art methods. 展开更多
关键词 single shot multibox DETECTOR (SSD) TV LOGO detection TV LOGO DATASET LOSS function HARD example mining
在线阅读 下载PDF
大规模连续中国手语数据集的创建与分析 预览
2
作者 袁甜甜 赵伟 +1 位作者 杨学 胡彬 《计算机工程与应用》 CSCD 北大核心 2019年第11期110-116,共7页
绝大多数健听人不懂手语导致听障人在找工作、就医、法律咨询等各生活、工作领域中遇到了极大的沟通障碍,而手语翻译员需要提前预约,成本也非常高,所以很多科研工作者都开始利用机器学习来开发手语自动翻译器,但其中的大部分研究都因为... 绝大多数健听人不懂手语导致听障人在找工作、就医、法律咨询等各生活、工作领域中遇到了极大的沟通障碍,而手语翻译员需要提前预约,成本也非常高,所以很多科研工作者都开始利用机器学习来开发手语自动翻译器,但其中的大部分研究都因为受到了数据集规模和质量的影响而效果不佳。为解决上述矛盾和问题,创建了目前全球最大的中国连续手语数据集,并使用了考虑身体关节的位置、面部表情及手指关节的端到端的深度学习模型进行有效训练。结论突显了现代深度学习技术在识别复杂手语方面的巨大优势,针对较小子集的BLEU-4已达到30.8。 展开更多
关键词 手语识别 深度学习 数据集 特征提取 端到端
在线阅读 下载PDF
Application of Big Data analysis in gastrointestinal research 预览
3
作者 Ka-Shing Cheung Wai K Leung Wai-Kay Seto 《世界胃肠病学杂志:英文版》 SCIE CAS 2019年第24期2990-3008,共19页
Big Data,which are characterized by certain unique traits like volume,velocity and value,have revolutionized the research of multiple fields including medicine.Big Data in health care are defined as large datasets tha... Big Data,which are characterized by certain unique traits like volume,velocity and value,have revolutionized the research of multiple fields including medicine.Big Data in health care are defined as large datasets that are collected routinely or automatically,and stored electronically.With the rapidly expanding volume of health data collection,it is envisioned that the Big Data approach can improve not only individual health,but also the performance of health care systems.The application of Big Data analysis in the field of gastroenterology and hepatology research has also opened new research approaches.While it retains most of the advantages and avoids some of the disadvantages of traditional observational studies(case-control and prospective cohort studies),it allows for phenomapping of disease heterogeneity,enhancement of drug safety,as well as development of precision medicine,prediction models and personalized treatment.Unlike randomized controlled trials,it reflects the real-world situation and studies patients who are often under-represented in randomized controlled trials.However,residual and/or unmeasured confounding remains a major concern,which requires meticulous study design and various statistical adjustment methods.Other potential drawbacks include data validity,missing data,incomplete data capture due to the unavailability of diagnosis codes for certain clinical situations,and individual privacy.With continuous technological advances,some of the current limitations with Big Data may be further minimized.This review will illustrate the use of Big Data research on gastrointestinal and liver diseases using recently published examples. 展开更多
关键词 Healthcare DATASET EPIDEMIOLOGY Gastric CANCER Inflammatory BOWEL disease Colorectal CANCER Hepatocellular carcinoma Gastrointestinal BLEEDING
在线阅读 免费下载
A Large Chinese Text Dataset in the Wild
4
作者 Tai-Ling Yuan Zhe Zhu +3 位作者 Kun Xu Cheng-Jun Li Tai-Jiang Mu Shi-Min Hu 《计算机科学技术学报:英文版》 SCIE EI CSCD 2019年第3期509-521,共13页
In this paper,we introduce a very large Chinese text dataset,in the wild.While optical character recognition(OCR)in document images is well studied and many commercial tools are available,the detection and recognition... In this paper,we introduce a very large Chinese text dataset,in the wild.While optical character recognition(OCR)in document images is well studied and many commercial tools are available,the detection and recognition of text in natural images is still a challenging problem,especially for some more complicated character sets such as Chinese text.Lack of training data has always been a problem,especially for deep learning methods which require massive training data.In this paper,we provide details of a newly created dataset of Chinese text with about 1 million Chinese characters from 3850 unique ones annotated by experts in over 30 000 street view images.This is a challenging dataset with good diversity containing planar text,raised text,text under poor illumination,distant text,partially occluded text,etc.For each character,the annotation includes its underlying character,bounding box,and six attributes.The attributes indicate the charactcr's background complexity,appearance,style,etc.Besides the dataset,we give baseline results using state-of-the-art methods for tliree tasks:character recognition(top-1 accuracy of 80.5%),character detection(AP of 70.9%),and text line detection(AED of 22.1).The dataset,source code,and trained models are publicly available. 展开更多
关键词 CHINESE TEXT DATASET CHINESE TEXT detection CHINESE TEXT RECOGNITION
特征选择方法在信用评分系统中的应用 预览
5
作者 吴锦华 王志生 +1 位作者 刘重阳 胡龙彪 《信息与电脑》 2019年第8期119-120,共2页
信用评分系统是在信用风险管理中比较重要的应用,可通过大数据分析技术构建评估分析模型来解决信用风险预测问题。具体而言:基于scikit-learn平台,利用平台中的特征选择方法构建有效模型,并将模型应用至实际数据集中得出信用评分,根据... 信用评分系统是在信用风险管理中比较重要的应用,可通过大数据分析技术构建评估分析模型来解决信用风险预测问题。具体而言:基于scikit-learn平台,利用平台中的特征选择方法构建有效模型,并将模型应用至实际数据集中得出信用评分,根据所得的评分结果向信用评估人员提供决策建议,从而降低最终风险。 展开更多
关键词 信用评分 scikit-learn 特征选择 数据集
在线阅读 下载PDF
反腐败开放政府数据建设研究 预览
6
作者 郭牧原 《情报探索》 2019年第6期96-100,共5页
[目的/意义]反腐败开放政府数据建设研究有助于在反腐败实践中更好地发挥其作用。[方法/过程]在总结现有相关研究工作的基础上,从关键环节和保障措施2个方面讨论反腐败开放政府数据的建设问题。[结果/结论]反腐败开放政府数据可在腐败... [目的/意义]反腐败开放政府数据建设研究有助于在反腐败实践中更好地发挥其作用。[方法/过程]在总结现有相关研究工作的基础上,从关键环节和保障措施2个方面讨论反腐败开放政府数据的建设问题。[结果/结论]反腐败开放政府数据可在腐败的预防、发现、调查、惩戒等环节发挥作用,在数据建设过程中需要重点解决数据集的确定、数据标准和发布等级的选择,以及政策保障、资金保障和组织保障等问题。 展开更多
关键词 反腐败 开放政府数据 数据集
在线阅读 下载PDF
数据集在人工智能医疗器械质控中的角色与要求 预览
7
作者 王浩 孟祥峰 +1 位作者 李澍 任海萍 《中国医疗器械杂志》 2019年第1期54-57,共4页
人工智能(artificial intelligence,AI)医疗器械是医疗器械新发展方向。其研发与质控需要高质量的临床数据集。由于人工智能医疗器械在国内外尚无标准规范,如何科学合理地构建数据集,如何发挥数据集的价值、降低临床试验成本,是产业发... 人工智能(artificial intelligence,AI)医疗器械是医疗器械新发展方向。其研发与质控需要高质量的临床数据集。由于人工智能医疗器械在国内外尚无标准规范,如何科学合理地构建数据集,如何发挥数据集的价值、降低临床试验成本,是产业发展的关键问题。该文参考国外行业和监管领域的现状和指导原则,分析了数据集在人工智能医疗器械质控中的角色与要求,对于监管部门制订人工智能医疗器械的监管决策提供支持,为全社会开发利用医疗数据提供参考。 展开更多
关键词 医疗器械 质量控制 人工智能 数据集
在线阅读 下载PDF
基于深度学习的水面漂浮物目标检测评估 预览
8
作者 雷李义 艾矫燕 +1 位作者 彭婧 姚冬宜 《环境与发展》 2019年第6期117-120,123共5页
在这篇文章中,我们提出了一个关于水面漂浮物的小型数据集,并分析了几种目标检测模型在数据集上的表现,包括FasterR-CNN,R-FCN和SSD。我们的目的是探究目标检测模型在检测水面漂浮物特别是非物体类别时的特性,并找出权衡精确度和速度后... 在这篇文章中,我们提出了一个关于水面漂浮物的小型数据集,并分析了几种目标检测模型在数据集上的表现,包括FasterR-CNN,R-FCN和SSD。我们的目的是探究目标检测模型在检测水面漂浮物特别是非物体类别时的特性,并找出权衡精确度和速度后最适合于引导水面清洁无人船的模型。为此,我们制作了一个小型的水面漂浮物数据集,数据集主要包括漂浮水草和漂浮落叶。之后我们通过将预训练模型在水面漂浮物数据集上进行迁移学习,实现了对于水面漂浮物区域的目标检测。我们对比并分析了这些模型的表现,SSD目标检测模型有着更高的精确度,FasterR-CNN模型则能给出更详细的预测,而同时拥有丰富结构特征和相当深度特征的模型对于困难目标有着更好的表现。 展开更多
关键词 数据集 深度学习 目标检测
在线阅读 下载PDF
Evaluation of the Forecast Performance for North Atlantic Oscillation Onset 预览
9
作者 Guokun DAI Mu MU Zhina JIANG 《大气科学进展:英文版》 SCIE CAS CSCD 2019年第7期753-765,共13页
By utilizing operational forecast products from TIGGE(The International Grand Global Ensemble) during 2006 to 2015,the forecasting performances of the European Centre for Medium-Range Weather Forecasts(ECMWF), Nationa... By utilizing operational forecast products from TIGGE(The International Grand Global Ensemble) during 2006 to 2015,the forecasting performances of the European Centre for Medium-Range Weather Forecasts(ECMWF), National Centers for Environmental Prediction(NCEP), Japan Meteorology Agency(JMA) and China Meteorological Administration(CMA) for the onset of North Atlantic Oscillation(NAO) events are assessed against daily NCEP–NCAR reanalysis data. Twenty-two positive NAO(NAO+) and nine negative NAO(NAO-) events are identified during this time period. For these NAO events,control forecasts, one member of the ensemble that utilizes the currently most proper estimate of the analysis field and the best description of the model physics, are able to predict their onsets three to five days in advance. Moreover, the failure proportion for the prediction of NAO-onset is higher than that for NAO+ onset, which indicates that NAO-onset is harder to forecast. Among these four operational centers, ECMWF has performs best in predicting NAO onset, followed by NCEP,JMA, and then CMA.The forecasting performance of the ensemble mean is also investigated. It is found that, compared with the control forecast, the ensemble mean does not improve the forecasting skill with respect to the onset time of NAO events. Therefore,a confident forecast of NAO onset can only be achieved three to five days in advance. 展开更多
关键词 NAO ONSET OPERATIONAL FORECAST TIGGE DATASET
在线阅读 下载PDF
软件开发活动数据集的层次化、多版本化方法 预览
10
作者 朱家鑫 周明辉 《软件学报》 EI CSCD 北大核心 2019年第7期2109-2123,共15页
随着开源软件的兴起及软件开发支撑工具的普及,Internet上积累了大量开放的软件开发活动数据,越来越多的实践者与研究者尝试从中获取提高软件开发效率和产品质量的洞察。为了提高数据分析的效率、方便分析结果的重现与对比,许多工作提... 随着开源软件的兴起及软件开发支撑工具的普及,Internet上积累了大量开放的软件开发活动数据,越来越多的实践者与研究者尝试从中获取提高软件开发效率和产品质量的洞察。为了提高数据分析的效率、方便分析结果的重现与对比,许多工作提出了构建与使用共享数据集。然而,现有软件开发活动数据集的构建过程可追溯性差、适用范围窄,对数据随时间、环境发生的变化欠考虑。这些不足直接威胁数据的质量及分析结果的有效性。针对该问题,提出一种层次化、多版本化的方法来构建与使用软件开发活动数据集。层次化是指在数据集中包括收集和后续处理所得的原始、中间和最终数据,建立数据集的可追溯性并扩展其适用范围。多版本化是指通过多种方式进行多次数据收集,使数据使用者能够观察到数据的变化,为数据质量及分析结果有效性的验证和提高创造条件。通过基于该方法构建的Mozilla问题追踪数据集进行示范,并验证了该方法能够帮助数据使用者高效地使用数据。 展开更多
关键词 数据驱动的软件工程 软件开发活动数据 数据分析 数据质量 数据集
在线阅读 下载PDF
基于深度学习的中国手语翻译 预览
11
作者 袁甜甜 胡彬 +1 位作者 杨学 赵伟 《电视技术》 2019年第2期52-55,80共5页
本文提出了一种“sequence-to-sequence”的方法来解决中国手语的翻译问题。基本模型包括用于视频特征提取的预训练CNN和用于脱机机器翻译的两层LSTM。为了提高性能,我们还引入一些开源工具用于捕获身体姿态并注释来作为额外的特征。同... 本文提出了一种“sequence-to-sequence”的方法来解决中国手语的翻译问题。基本模型包括用于视频特征提取的预训练CNN和用于脱机机器翻译的两层LSTM。为了提高性能,我们还引入一些开源工具用于捕获身体姿态并注释来作为额外的特征。同时,使用Kinect2.0采集的大量样本作为训练数据,并提出一种面向手语者的时空注意力模型,以提高翻译的准确性。为了惠及其他研究人员和聋人,并帮助推广中国通用手语,我们已建立了一个较为庞大的连续中文手语语料库,并准备在全球范围内进行部分数据的共享。 展开更多
关键词 手语识别 神经机器翻译 CNN 时空注意力模型 语料库
在线阅读 下载PDF
Low microRNA-139 expression associates with poor prognosis in patients with tumors:A meta-analysis 预览
12
作者 Jian-An Chen Yan Yu +6 位作者 Chen Xue Xiao-Long Chen Guang-Ying Cui Juan Li Kong-Fei Li Zhi-Gang Ren Ran-Ran Sun 《国际肝胆胰疾病杂志:英文版》 SCIE CAS CSCD 2019年第4期321-331,共11页
Background:microRNA-139(miR-139)is dysregulated in various types of tumors and plays a key role in carcinogenesis.miR-139 may be used as a diagnostic and prognostic biomarker of cancers.However,the data from the liter... Background:microRNA-139(miR-139)is dysregulated in various types of tumors and plays a key role in carcinogenesis.miR-139 may be used as a diagnostic and prognostic biomarker of cancers.However,the data from the literature are not consistent.The present study aimed to verify the prognostic and diagnostic values of miR-139 in solid tumors.Data sources:PubMed,Web of Science and Embase databases were searched and publications from January 2011 to August 2017 were included.We used Gene Expression Omnibus(GEO)and The Cancer Genome Atlas(TCGA)database to further validate this meta-analysis.Results:Eight individual studies from seven articles were included.Pooled analyses showed that low miR-139 expression was related to worse overall survival(OS)[hazard ratio(HR)=2.27;95%confidence intervals(CI):1.74–2.95;P<0.001]in solid tumors,including hepatocellular carcinoma(HCC)and glioblastoma multiforme(GBM),consisting with the results of TCGA.However,our results of CRC showed that low miR-139 expression was associated with poor OS which was contradictory with the results in TCGA database and need larger samples to validate the phenomenon;whereas for CRC patients,high miR-139 expression predicted poor RFS,which was in good accordance with TCGA results.The results of 27 microarrays from GEO database showed that miR-139 expression levels were lower in tumor tissues compared to adjacent non-tumor tissues or healthy tissues.Decreased miR-139 expression was also significantly correlated with poor differentiation grade(OR=3.57;95%CI:1.44–8.85;P=0.006).However,the combined data indicated that no associations between miR-139 expression and the following parameters such as age(pooled OR=1.50;95%CI:0.69–3.24;P=0.304),gender(pooled OR=0.92;95%CI:0.56–1.51;P=0.738),tumor size(pooled OR=1.51;95%CI:0.69–3.31;P=0.298),late tumor-node-metastasis stage(pooled OR=1.63;95%CI:0.99–2.68;P=0.057)and lymph-node-metastasis(pooled OR=0.66;95%CI:0.34–1.28;P=0.222).Conclusions:Low miR-139 expression was related to poor prognosis in HCC and 展开更多
关键词 MICRORNAS microRNA-139 TCGA dataset GEO database PROGNOSIS
在线阅读 下载PDF
基于卷积神经网络的文本检测算法研究 预览
13
作者 李阳 李绍彬 +1 位作者 解云超 冯爽 《中国传媒大学学报:自然科学版》 2019年第1期70-76,共7页
随着深度学习的发展,利用神经网络对文本进行检测得到了更深入的研究和更广泛的应用。本文基于Text-Boxes算法,在考虑到足球赛事场景下的文本特点后对其进行改进,提升了该场景下文本检测的效果。针对足球赛事场景下的文本几何形状多样... 随着深度学习的发展,利用神经网络对文本进行检测得到了更深入的研究和更广泛的应用。本文基于Text-Boxes算法,在考虑到足球赛事场景下的文本特点后对其进行改进,提升了该场景下文本检测的效果。针对足球赛事场景下的文本几何形状多样性的特点,设置适应于足球场景中文本检测的默认框;针对影响模型优化的样本不均衡问题,使用Focal Loss作为用于分类的损失函数;最后使用非极大值抑制过滤冗余的矩形框,获得最终的检测结果。本文自行标注了足球赛事场景数据集,用于网络的训练,验证了本文算法的有效性。 展开更多
关键词 文本检测 卷积神经网络 数据集
在线阅读 下载PDF
Deep learning for in vitro prediction of pharmaceutical formulations
14
作者 Yilong Yang Zhuyifan Ye +3 位作者 Yan Su Qianqian Zhao Xiaoshan Li Defang Ouyang 《药学学报:英文版》 CSCD 2019年第1期177-185,共9页
Current pharmaceutical formulation development still strongly relies on the traditional trialand-error methods of pharmaceutical scientists. This approach is laborious, time-consuming and costly.Recently, deep learnin... Current pharmaceutical formulation development still strongly relies on the traditional trialand-error methods of pharmaceutical scientists. This approach is laborious, time-consuming and costly.Recently, deep learning has been widely applied in many challenging domains because of its important capability of automatic feature extraction. The aim of the present research is to apply deep learning methods to predict pharmaceutical formulations. In this paper, two types of dosage forms were chosen as model systems. Evaluation criteria suitable for pharmaceutics were applied to assess the performance of the models. Moreover, an automatic dataset selection algorithm was developed for selecting the representative data as validation and test datasets. Six machine learning methods were compared with deep learning. Results showed that the accuracies of both two deep neural networks were above 80% and higher than other machine learning models;the latter showed good prediction of pharmaceutical formulations. In summary, deep learning employing an automatic data splitting algorithm and the evaluation criteria suitable for pharmaceutical formulation data was developed for the prediction of pharmaceutical formulations for the first time. The cross-disciplinary integration of pharmaceutics and artificial intelligence may shift the paradigm of pharmaceutical research from experience-dependent studies to data-driven methodologies. 展开更多
关键词 PHARMACEUTICAL FORMULATION Deep learning Small data Automatic DATASET selection algorithm ORAL fast disintegrating films ORAL SUSTAINED release matrix TABLETS
Forest carbon storage in Guizhou Province based on field measurement dataset
15
作者 Chunzi Guo Yangyang Wu +1 位作者 Jian Ni Yinming GUO 《中国地球化学学报:英文版》 EI CAS CSCD 2019年第1期8-21,共14页
Accurate estimation of forest carbon storage is crucial in understanding global and regional carbon cycles and projecting future ecological and economic scenarios.Guizhou is the largest karst landform province in Chin... Accurate estimation of forest carbon storage is crucial in understanding global and regional carbon cycles and projecting future ecological and economic scenarios.Guizhou is the largest karst landform province in China;61.9%of its land area is characterized as karst.However,monitoring its field biomass and carbon storage is difficult.This study synthesized and analyzed a comprehensive database of direct field observations of forest vegetation and soil carbon storage in Guizhou Province by using data from existing literature.The total vegetation carbon storage in Guizhou Province was 488.170 TgC,the average vegetation carbon density (VCD)was 27.866 MgC hm^-2,the total amount of soil organic carbon (SOC)(20 cm)was 1017.364 TgC,and the average SOC density was 58.074 MgC hm^-2.Among all vegetation types,needleleaf forest had the highest vegetation carbon stocks,and scrub presented the highest SOC storage.The vegetation and SOC storage values of the karst landform were 282.352 and 614.825 TgC,respectively,which were higher than those of the non-karst landform.VCD was concentrated at 10M0 MgC hm^-2,and SOC density was concentrated at 40-60,60-80,and 80-100 MgC hm^-2.This comprehensive regional data synthesis and analysis based on direct field measurement of vegetation and soil will improve our understanding of the forest carbon cycle in karst landforms under a changing climate. 展开更多
关键词 FOREST carbon STORAGE Field measurement DATASET KARST LANDFORM
医院人力资源管理基本数据集及数据元标准研究 预览
16
作者 刘微 刘建超 +1 位作者 冯丹 刘丽华 《中国医院》 2019年第5期70-72,共3页
目的:为规范基于企业资源管理(enterpriseresourceplanning,ERP)的人力资源管理信息系统建设,促进医院人、财、物运营数据的交换与共享,研究建立医院人力资源管理基本数据集标准。方法:围绕现代医院人力资源管理的核心内容,参照国内外... 目的:为规范基于企业资源管理(enterpriseresourceplanning,ERP)的人力资源管理信息系统建设,促进医院人、财、物运营数据的交换与共享,研究建立医院人力资源管理基本数据集标准。方法:围绕现代医院人力资源管理的核心内容,参照国内外较为成熟的医院ERP人力资源运营管理信息系统,应用数据集与数据元标准化技术和《WS363.1-2011卫生信息数据元目录第1部分:总则》发布的描述规范,构建国内通用性的人力资源运营管理信息系统基本数据集,对数据集相应数据元属性进行规范化描述。结果:建立了7个主题数据集,梳理出48个通用数据元并提交了规范化描述结果,定义了11个数据元值域代码表,形成国内基于医院资源管理的《医院人力资源管理基本数据集》卫生行业标准,标准号:WS599.1-2018。结论:医院人力资源管理基本数据集标准是医院人、财、物信息系统建设的基础,是促进医院间人力资源信息系统互联互通和数据共享的条件。 展开更多
关键词 医院 人力资源管理 数据集 数据元 标准
在线阅读 下载PDF
Open-source dataset for control-oriented modelling in diesel engines
17
作者 Jinghua ZHAO Sitong ZHOU +3 位作者 Yunfeng HU Mingjun JU Ruixue REN Hong CHEN 《中国科学:信息科学(英文版)》 SCIE EI CSCD 2019年第7期223-224,共2页
In this study, an experimental dataset of a diesel engine under the European transient cycle (ETC)is demonstrated and opened for academic research.The dataset with 7 dimensional features is collected from the advanced... In this study, an experimental dataset of a diesel engine under the European transient cycle (ETC)is demonstrated and opened for academic research.The dataset with 7 dimensional features is collected from the advanced diesel engine manufactured by Changchun FAW Sihuan Engine Manufacture Co., Ltd. 展开更多
关键词 control-oriented MODELLING EXPERIMENTAL DATASET engine
A Novel Dataset For Intelligent Indoor Object Detection Systems 预览
18
作者 Mouna Afif Riadh Ayachi +2 位作者 Yahia Said Edwige Pissaloux Mohamed Atri 《人工智能进展(英文)》 2019年第1期52-58,共7页
Indoor Scene understanding and indoor objects detection is a complex high-level task for automated systems applied to natural environments.Indeed,such a task requires huge annotated indoor images to train and test int... Indoor Scene understanding and indoor objects detection is a complex high-level task for automated systems applied to natural environments.Indeed,such a task requires huge annotated indoor images to train and test intelligent computer vision applications.One of the challenging questions is to adopt and to enhance technologies to assist indoor navigation for visually impaired people(VIP)and thus improve their daily life quality.This paper presents a new labeled indoor object dataset elaborated with a goal of indoor object detection(useful for indoor localization and navigation tasks).This dataset consists of 8000 indoor images containing 16 different indoor landmark objects and classes.The originality of the annotations comes from two new facts taken into account:(1)the spatial relationships between objects present in the scene and(2)actions possible to apply to those objects(relationships between VIP and an object).This collected dataset presents many specifications and strengths as it presents various data under various lighting conditions and complex image background to ensure more robustness when training and testing objects detectors.The proposed dataset,ready for use,provides 16 vital indoor object classes in order to contribute for indoor assistance navigation for VIP. 展开更多
关键词 INDOOR OBJECT detection and recognition INDOOR image DATASET Visually IMPAIRED People(VIP) Idoor NAVIGATION
在线阅读 免费下载
2018机器阅读理解技术竞赛总体报告 预览
19
作者 刘凯 刘璐 +4 位作者 刘璟 吕雅娟 佘俏俏 张倩 时迎超 《中文信息学报》 CSCD 北大核心 2018年第10期118-129,共12页
机器阅读理解是自然语言处理和人工智能领域的前沿课题,"2018机器阅读理解技术竞赛"旨在推动相关技术研究和应用的发展。竞赛发布了最大规模的中文阅读理解数据集,提供了先进的开源基线系统,采用改进的自动评价指标,吸引了国内外千余... 机器阅读理解是自然语言处理和人工智能领域的前沿课题,"2018机器阅读理解技术竞赛"旨在推动相关技术研究和应用的发展。竞赛发布了最大规模的中文阅读理解数据集,提供了先进的开源基线系统,采用改进的自动评价指标,吸引了国内外千余支队伍参与,参赛系统效果提升显著。该文详细介绍技术竞赛的总体情况、竞赛设置、组织流程、评价结果,并对参赛系统结果进行了分析。 展开更多
关键词 机器阅读理解 自动问答 数据集 技术评测
在线阅读 下载PDF
MHW蒙古文脱机手写数据库及其应用 预览
20
作者 范道尔吉 高光来 武慧娟 《中文信息学报》 CSCD 北大核心 2018年第1期89-95,共7页
建立公开、权威的蒙古文手写数据库是研究和开发蒙古文手写识别系统的基础。该文在蒙古文编码、构词和语法的研究基础上,公开了一个蒙古文大词汇量脱机手写数据库MHW,其中训练集由5 000个单词构成,每个词采集了20个样本,共包含10万样本... 建立公开、权威的蒙古文手写数据库是研究和开发蒙古文手写识别系统的基础。该文在蒙古文编码、构词和语法的研究基础上,公开了一个蒙古文大词汇量脱机手写数据库MHW,其中训练集由5 000个单词构成,每个词采集了20个样本,共包含10万样本,测试集Ⅰ包含5 000样本,测试集Ⅱ包含14 085样本。该文利用蒙古文文字长度可变特征研究了自动错误检测算法,提高了字库的可靠性。在三种常用手写识别模型上评估了字库的性能,其中基于循环神经网络的模型表现出最佳性能,在字典受限条件下测试集Ⅰ的词错误率达到2.20%,测试集Ⅱ达到了5.55%。 展开更多
关键词 蒙古文 手写识别 字库 HMM LSTM
在线阅读 下载PDF
上一页 1 2 47 下一页 到第
使用帮助 返回顶部 意见反馈