您当前所在位置: 首页 > 首发论文
筛选条件

时间

领域

全部

计算机科学技术(13)

信息科学与系统科学(1)

电子、通信与自动控制技术(1) 显示更多>>

中医学与中药学(0) 临床医学(0) 交通运输工程(0) 体育科学(0) 农学(0) 冶金工程技术(0) 力学(0) 动力与电气工程(0) 化学(0) 化学工程(0) 图书馆、情报与文献学(0) 土木建筑工程(0) 地球科学(0) 基础医学(0) 天文学(0) 安全科学技术(0) 工程与技术科学基础学科(0) 心理学(0) 教育学(0) 数学(0) 机械工程(0) 材料科学(0) 林学(0) 核科学技术(0) 水产学(0) 水利工程(0) 测绘科学技术(0) 物理学(0) 环境科学技术(0) 生物学(0) 畜牧科学、动物医学(0) 矿山工程技术(0) 管理学(0) 纺织科学技术(0) 经济学(0) 能源科学技术(0) 航空航天科学技术(0) 药学(0) 预防医学与卫生学(0) 食品科学技术(0)

学术评议

实时热搜榜

我的筛选 >
2003-2020 全部
为您找到包含“深度卷积神经网络”的内容共15

赵钰,董远,白洪亮

2017-11-20

近年来,深度卷积神经网络在目标分类、物体检测等任务中均获得了很突出的结果。基于深度卷积神经网络的人脸识别也在公开数据集中超过了人眼的识别能力。在人脸特征提取的训练中,特征嵌入已经成为了比较热门的方法

国家自然科学基金重点项目(61532018

北京邮电大学信息与通信工程学院,北京 100876 , 北京邮电大学信息与通信工程学院,北京 100876 , 北京飞搜科技有限公司,北京 100081

#计算机科学技术#

1评论(0 分享(0)

Wang Shuainan,Xu Tong,Li Wei,Sun Haifeng

Face detection has achieved great success with the development of convolution neural network. However, it remains a challenging problem to detect small and blurred faces in unconstrained environment. This paper presents a novel cascade single-shot face detector, named Cascade Single Shot Face Detector (CSSD), which introduces novel cascade classification and regression network in an anchor-based face detector to reject false positives and improve location accuracy. We have contributed in the following three aspects: 1) proposing a feature enchanted and scale-invariable face detection architecture to process faces with different scales; 2) regressing bounding boxes of faces in two steps with a cascade method; 3) filtering negative anchors online after anchor refinement and rebalancing puzzle negative anchors and positive anchors with rate of 3:1. As a consequence, our method achieves state-of-the-art detection performance on FDDB and WIDER FACE dataset.

2019-01-25

State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications;State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications;State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications;State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications,,,

#Computer Science and Technology#

0评论(0 分享(0)

ZHANG Jian-Wei,GUO Qiu-Shan,DONG Yuan,XIONG Feng-Ye,BAI Hong-Liang

Face recognition has achieved great success due to the development of Deep convolutional neural networks (DCNN). Loss functions with angular margin have been proposed to supervise DCNN for better feature representation. However, these methods would suffer from sensitivity of hyperparameters setting. In this paper, we propose an Adaptive Parameters Softmax Loss function with different scale parameters for target logits and non-target logits and dynamically adaptive margin parameter. Extensive experiments on MegaFace and IJB-C demonstrate the effectiveness of our method.

2019-09-24

National Natural Science Foundation (61532018

School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876,School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876,School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876,Beijing FaceALL Technology Ltd., Beijing 100081,Beijing FaceALL Technology Ltd., Beijing 100081

#Computer Science and Technology#

0评论(0 分享(0)

易鸣,罗涛

2016-01-12

本文针对现有基于深度学习的人脸特征点定位算法对于头部姿态变化鲁棒性较差的问题,提出了一种基于多任务深度卷积神经网络(Multi-task Deep Convolutional Neural

北京邮电大学信息与通信工程学院, 北京 100876,北京邮电大学信息与通信工程学院, 北京 100876

#计算机科学技术#

0评论(0 分享(0)

马震,杨辉华,潘细朋

2019-03-20

皮肤科医生的工作量,提高诊断效率和诊断的客观性,从而帮助患者及早确诊治疗。近年来,深度卷积神经网络因其强大的自学习特征提取能力,在图像识别、分割、检测等领域得到广泛研究与应用。因此,本文研究基于深度卷积

Automation School, Beijing University of Posts and Telecommunications, Beijing 100876,Automation School, Beijing University of Posts and Telecommunications, Beijing 100876;Computer and Information Security School, Guilin University of Electronic Technology, Guilin 541004,Automation School, Beijing University of Posts and Telecommunications, Beijing 100876

#计算机科学技术#

0评论(0 分享(0)

LIU Yu-xuan, DONG Yuan, BAI Hong-liang

Scene Classification is a subdivision problem of Large-scale classfication problem since the latter has been basically resolved. In this article, several common Scene Classification Data-set and their differences are introduced. Additionally, there are lots of advanced methods of Deep Convolutional Neural Network. Methods for solving Large-scale Classification problems to be used on solving Scene Classification is a very common way. This article summerizes the results of those network structures trained on Scene Data-sets. Therefore, this article introduces some improvement for simply using CNN on Scene Classification and their better result. Since the common network structure is so complicated that it takes a long time to train and test, a method of simplifying these deep networks is raised in this article. Reducing size of input pictures and numbers of convolution kernels could take effect on increasing the speed on both training and testing stages. Finally, this much smaller network got an acceptable result on the data-set. % Reviews: please describe the background, status and application of the research with 150-300 words. I and we can not be used as the subject, % and the abstract must not the same as the sentences of the main text. General research paper: please extracts the key points of the paper, give the main research achievements with object, methods, results and conclusion with 200-400 words. I and we can not be used as the subject, and the abstract must not the same as the sentences of the main text.

2016-07-05

School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876 , School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876 , Beijing FaceALL Technology Ltd., Beijing 100082

#Electrics, Communication and Autocontrol Technology#

刘娜,李翠华

2015-03-24

;在记忆阶段使用深度卷积神经网络学习得到超分辨率重建的过完备字典;在决策阶段,将过完备字典作为超分辨率重建的依据,对单帧图像进行超分辨率重建。实验结果表明,该模型对单帧图像具有良好的重建能力,能够较好地

高等学校博士学科点专项科研基金(20110121110020

国家自然科学基金(61373077

厦门大学信息科学与技术学院计算机科学系,厦门 361005,厦门大学信息科学与技术学院计算机科学系,厦门 361005

#计算机科学技术#

本文收录在中国科技论文,2015,10(2):201-206.

0评论(0 分享(0)

XIAO Degui,ZHONG Pei

Deep convolutional neural networks (DCNNs) have recently demonstrated state-of-the-art performance in advanced vision tasks, such as image classification and object detection. This work focuses on solving image semantic segmentation tasks. First,we combine a new feature extraction network with adilated convolution layer to improve the accuracy of the model's mission. Second, we introduce multi-scale feature fusion technology to improve the performance of DCNN. Third, we combine the DCNN with fully connected conditional random field to overcome the inaccurate positioning of DCNN and optimize their output. Our approach is demonstrated on the PASCAL VOC-2012 Image Semantic Segmentation dataset, where 78.1% IOU accuracy is achieved in the test set. Our approach can compute neural network responses intensively at 9 frames per second on modern GPUs.

2019-04-04

National Natural Science Foundation of China (61272062

Hunan University, College of Computer Science and Electronic Engineering, Changsha 410006,Hunan University, College of Computer Science and Electronic Engineering, Changsha 410006

#Computer Science and Technology#

0评论(0 分享(0)

Guanghua Tan,Rui Miao

Recently, deep convolutional neural networks have made great breakthroughs in the field of action recognition. Since sequential video frames have a lot of redundant information, compared with dense sampling, sparse sampling network can also achieve good results. Due to sparse sampling\'s limitation of access to information, this paper mainly discusses how to further improve the learning ability of the model based on sparse sampling. We proposed a model based on divide-and-conquer, which use a threshold α to determine whether action data require sparse sampling or dense local sampling for learning. Finally, our approach obtains the state-the-of-art performance on the datasets of HMDB51 (72.4%) and UCF101 (95.3%).

2019-04-25

Natural Science Foundation of Hunan Province (2018JJ3074

Hunan University,College of Computer Science and Electronic Enginee,Changsha 410000,Hunan University,College of Computer Science and Electronic Enginee,Changsha 410000

#Computer Science and Technology#

0评论(0 分享(0)

韩凝,高升

2018-05-21

卷积神经网络算法,分别探讨了歌词文本、音频信号和二者结合作为输入信息时音乐自动标注模型的构建。在自主收集构建的中文音乐标注数据集上,分析比较了基于字向量的深度神经网络自动标注模型、基于音频的卷积神经网络

School of Information and Communication Engineering, Beijing University of Posts and Telecommunications,School of Information and Communication Engineering, Beijing University of Posts and Telecommunications

#计算机科学技术#

0评论(0 分享(0)