检索结果-内蒙古大学图书馆

Multi-Modality Adversarial auto-encoder for Zero-Shot Learning

IEEE ACCESS 2020年 8卷 9287-9295页

作者： Ji, Zhong Dai, Guangwen Yu, Yunlong Tianjin Univ Sch Elect & Informat Engn Tianjin 300072 Peoples R China Zhejiang Univ Coll Informat Sci & Elect Engn Hangzhou 310027 Peoples R China

The existing generative Zero-Shot Learning (ZSL) methods only consider the unidirectional alignment from the class semantics to the visual features while ignoring the alignment from the visual features to the class semantics, which fails to construct the visual-semantic interactions well. In this paper, we propose to generate visual features based on an auto-encoder framework paired with multi-modality adversarial networks respectively for visual and semantic modalities to reinforce the visual-semantic interactions with a bidirectional alignment, which ensures the generated visual features to fit the real visual distribution and to be highly related to the semantics. The encoder aims at generating real-like visual features while the decoder forces both the real and the generated visual features to be more related to the class semantics. To further capture the discriminative information of the generated visual features, both the real and generated visual features are forced to be classified into the correct classes via a classification network. Experimental results on four benchmark datasets show that the proposed approach is particularly competitive on both the traditional ZSL and the generalized ZSL tasks.

关键词： Zero-shot learning adversarial network auto-encoder image recognition

来源：评论

学校读者我要写书评

暂无评论

An auto-encoder-based summarization algorithm for unstructured videos

引用

MULTIMEDIA TOOLS AND APPLICATIONS 2017年第23期76卷 25039-25056页

作者： Han, Meng-Xiong Hu, Hai-Miao Liu, Yang Zhang, Chi Tian, Rong-Peng Zheng, Jin Beihang Univ Sch Comp Sci & Engn Beijing Key Lab Digital Media Beijing 100083 Peoples R China Beijing Inst Graph Beijing 100029 Peoples R China Beihang Univ Digital Media Lab Beijing Peoples R China

Video summarization is an effective way to quick view videos and relieve the pressure of videos storage. However the traditional algorithms are hardly adapted to unstructured videos, due to the unobvious for scenes changing and ignoring the structure of the videos. Therefore, an auto-encoder-based summarization algorithm is proposed in this paper for unstructured videos. Each video structure is detected by an auto-encoder and both of the interestingness and representativeness of each video segment are predicted by the reconstruction errors of the segment. Meanwhile, most interesting and representative summarization is generated with the limited summary length. The experimental results show that the proposed algorithm obtained a better performance by comparing with the state-of-the-art.

关键词： Video summarization auto-encoder Video analysis

来源：评论

学校读者我要写书评

暂无评论

Fuzzy One-Class Extreme auto-encoder

引用

NEURAL PROCESSING LETTERS 2019年第1期50卷 701-727页

作者： Yu, Hualong Sun, Dan Xi, Xiaoyan Yang, Xibei Zheng, Shang Wang, Qi Jiangsu Univ Sci & Technol Sch Comp 2 Mengxi Rd Zhenjiang 212003 Jiangsu Peoples R China

A novel one class classification (OCC) algorithm called fuzzy one class extreme auto-encoder (FOCEAE) is presented in this article. The algorithm combines the precision of probability density estimation and the generalization of neural networks to accurately generate the compact bound for the target class cases. Firstly, a K-nearest-neighbors non-parametric probability density estimation-alike strategy is used to estimate the relative densities of all target class training objects, then the relative densities are transformed to be the fuzzy coefficients for further training fuzzy extreme learning machine (FELM) model. Specifically, considering there are only one-class instances, FELM is trained in the form of auto-encoder, i.e., each input equals to be the expected output of the network. Finally, the bound (i.e., the threshold) of the target class cases is determined by calculating and ranking the reconstructed errors of all training instances. We show the effectiveness and superiority of the proposed FOCEAE algorithm by comparing it with some benchmark OCC algorithms on a mass of data sets in terms of both F-measure and G-mean metrics. The statistical results also indicate that the proposed algorithm performs significantly better than some conventional ones.

关键词： One class classification Fuzzy extreme learning machine auto-encoder Probability density estimation K nearest neighbors Reconstruction error

来源：评论

学校读者我要写书评

暂无评论

Exemplar-based Cascaded Stacked auto-encoder Networks for robust face alignment

引用

COMPUTER VISION AND IMAGE UNDERSTANDING 2018年 171卷 95-103页

作者： Zhang Junfeng Hu Haifeng Sun Yat Sen Univ Sch Elect & Informat Engn Guangzhou Guangdong Peoples R China

In this paper, we present a novel Exemplar-based Cascaded Stacked auto-encoder Network (ECSAN) for facial landmarks detection. The proposed framework consists of a Global Exemplar Constraint Stacked auto-encoder Network (GECSAN) and a set of Local Information Preserve Stacked auto-encoder Networks (LIPSANs). In our work, GECSAN utilizes successive stacked auto-encoder network and some well-designed exemplars to obtain an initial shape estimation from a holistic facial image. Then LIPSANs are presented which take the local features extracted around current landmarks as input and generate a facial landmark refinement. Different from existing deep models, a prior exemplar-based shape is utilized to handle the partial occlusion in the image so that our model can achieve robustness against local occlusions. Experimental results on several datasets demonstrate that our model acquires better performance over the state-of-the-art methods with respect to occlusion handling and attain higher alignment accuracy.

关键词： auto-encoder Deep learning Exemplar-based Cascaded architecture Face alignment

来源：评论

学校读者我要写书评

暂无评论

Unsupervised deep learning framework for ultrasonic-based distributed damage detection in concrete: integration of a deep auto-encoder and Isolation Forest for anomaly detection

引用

STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL 2024年第3期23卷 1313-1333页

作者： Toufigh, Vahab Ranjbar, Iman Sharif Univ Technol Dept Civil Engn Tehran Iran Sharif Univ Technol Dept Civil Engn Room 428 4th Floor Civil Engn Bldg Azadi St Tehran 111551639 Iran

This study presented an unsupervised anomaly detection-based framework for distributed damage detection in concrete using ultrasonic response signals. A deep fully connected auto-encoder was employed to reconstruct the ultrasonic response signals. This model was trained on the intact specimen's responses. The auto-encoder demonstrated a relatively high prediction error encountering the damaged specimen's responses. Two time-domain features (mean squared error and reconstructed-to-original signal ratio) and one frequency-domain feature (fundamental amplitude ratio) were defined to measure the reconstruction error of the auto-encoder (the damage-sensitive features). Finally, the Isolation Forest algorithm was implemented for anomaly (damage) detection. The beauty of this framework is that it requires a few numbers of data only from the intact specimen for training the auto-encoder and collecting the binary decision trees of the Isolation Forest. The framework was successfully implemented for damage detection in five geopolymer concrete specimens with different mix proportions. Using all three introduced damage-sensitive features, the framework demonstrated an average prediction accuracy of 95.0% and 93.0% for damaged and intact stages, respectively.

关键词： Concrete damage detection ultrasonic unsupervised deep learning auto-encoder anomaly detection isolation forest

来源：评论

学校读者我要写书评

暂无评论

Cross -modal guidance based auto -encoder for multi -video summarization

引用

PATTERN RECOGNITION LETTERS 2020年 135卷 131-137页

作者： Ji, Zhong Zhao, Yuxiao Pang, Yanwei Li, Xuelong Tianjin Univ Sch Elect & Informat Engn Tianjin 300072 Peoples R China Northwestern Polytech Univ Ctr OPT IMagery Anal & Learning Xian 710072 Shaanxi Peoples R China

Multi-Video Summarization (MVS) aims at condensing a great number of videos with the same search query to obtain a compact storyboard, which improves user’s browsing experience. Recently, a line of research resorts to a mass of extra web images to specify the search intent for a query. However, these web images are indirect and noisy. In contrast, the tag information for queried videos is more accurate and concise than those web images, which can also be directly employed as side information. To this end, we proposed to employ tag information, i.e., titles and descriptions, as the side information for the generation of summarization. Specifically, we employed sparse auto-encoder as the main model to generate the final summarization, where the input was the multiple videos, the output was the keyframes set. Meanwhile, we fused the visual and tag information to guide visual features, which constrained the sparse auto-encoder to select the important candidate keyframes. The proposed approach is called Cross-modal guidance based auto-encoder for MVS approach (CAE-MVS). Extensive experimental results demonstrate that the CAE-MVS clearly outperforms the state-of-the-art methods.

关键词： Multi-video summarization auto-encoder Cross-modal

来源：评论

学校读者我要写书评

暂无评论

Attention-based misaligned spatiotemporal auto-encoder for video anomaly detection

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2024年第SUPPL 1期18卷 285-297页

作者： Yang, Haiyan Liu, Shuning Wu, Mingxuan Chen, Hongbin Zeng, Delu Guilin Univ Elect Technol Sch Informat & Commun Jinji Rd Guilin 541004 Peoples R China South China Univ Technol Sch Elect & Informat Engn Wushan Rd Guangzhou 510641 Peoples R China

To face the shortcomings of the auto-encoder (AE) algorithm in terms of weak dynamic object extraction and strong reconstruction ability in abnormal situations for video anomaly detection, an attention-based misaligned spatiotemporal AE (AMS-AE) model is proposed for video anomaly detection. The AMS-AE model uses a pseudo-abnormal generation module (PAGM) to dislocate the spatiotemporal information of the video and then uses an attention mechanism algorithm to selectively capture the foreground regions. Finally, it reconstructs the video frames through encoding and decoding. During the training process, normal data is used for training, while the abnormal frames generated by the PAGM are introduced as anomalies to limit the reconstruction ability of the auto-encoding algorithm on abnormal data, which facilitates better detection of anomalies during testing phase. The experimental results show that the AMS-AE algorithm achieves significantly better performance than some other algorithms on three public datasets (UCSD Ped2, CUHK Avenue, and Shanghai Tech), effectively improving the accuracy and robustness of video anomaly detection. The area under the ROC curve (AUC) values of 98.7%, 89.7%, and 74.0% were achieved on the above public datasets, respectively.

关键词： Video anomaly detection Pseudo-anomaly generation module Attention mechanism auto-encoder

来源：评论

学校读者我要写书评

暂无评论

Learning Cascaded Deep auto-encoder Networks for Face Alignment

引用

IEEE TRANSACTIONS ON MULTIMEDIA 2016年第10期18卷 2066-2077页

作者： Weng, Renliang Lu, Jiwen Tan, Yap-Peng Zhou, Jie Nanyang Technol Univ Sch Elect & Elect Engn Singapore 639798 Singapore Tsinghua Univ Dept Automat Beijing 100084 Peoples R China

In this paper, we propose a new cascaded deep auto-encoder networks (CDAN) approach for face alignment. Our framework consists of a global exemplar-based deep auto-encoder network (GEDAN) and a series of localized deep auto-encoder networks (LDAN) in a cascaded fashion. The global network takes a low-resolution holistic facial image as input and generates a preliminary facial landmark configuration. The following localized networks sample pose-indexed local features around current landmark positions, and refine the landmark positions with increasingly higher image resolutions. Our network architectures are designed to achieve greater robustness against pose variations as well as higher landmark estimation accuracy. Experimental results on three datasets show that the proposed approach achieves superior alignment accuracy with real-time speed.

关键词： auto-encoder deep learning face alignment

来源：评论

学校读者我要写书评

暂无评论

A feature extraction and classification algorithm based on improved sparse auto-encoder for round steel surface defects

引用

MATHEMATICAL BIOSCIENCES AND ENGINEERING 2020年第5期17卷 5369-5394页

作者： Yan, Xuguo Gao, Liang Huazhong Univ Sci & Technol Sch Mech Sci & Engn State Key Lab Digital Mfg Equipment & Technol Wuhan 430074 Peoples R China

Traditional feature dimensionality reduction (FDR) algorithms can extract features by reducing feature dimensions. However, it may lose some useful information and affect the accuracy of classification. Normally, in traditional defect feature extraction, it first obtain the defect area of the defect image by image preprocessing and defect segmentation, select the original feature set of defects by prior knowledge, and extract the optimal features by traditional FDR algorithms to solve the problem of "curse of dimensionality". In this paper, a feature extraction and classification algorithm based on improved sparse auto-encoder (AE) is proposed. We adopt three traditional FDR algorithms at the same time, combine the defect features obtained in pairs, take the merged defect features as the input of sparse AE, then use the "bottleneck" of sparse AE to conduct the defects classification by Softmax classifier. The experimental results show that the proposed algorithm can extract the optimal features of round steel surface defects with less network training time than individual sparse AE, finally get higher classification accuracy than individual FDR algorithm in the actual production line.

关键词： feature extraction feature dimensionality reduction auto-encoder defect detection round steel

来源：评论

学校读者我要写书评

暂无评论

Representation learning by hierarchical ELM auto-encoder with double random hidden layers

引用

IET COMPUTER VISION 2019年第4期13卷 411-419页

作者： Li, Rui Wang, Xiaodan Lei, Lei Wu, Chongming Air Force Engn Univ Coll Air & Missile Def Xian 710051 Shanxi Peoples R China Xijing Univ Coll Business Xian 710123 Shanxi Peoples R China

Recent research developments of extreme learning machine (ELM) with multilayer network architecture lead to a promising high performance with extremely fast training speed for representation learning. In this work, the authors are dedicated to develop an efficient and expressive representation learning method with hierarchical ELM, and proposing a novel architectural unit named as double random hidden layers ELM auto-encoder (DELM-AE). The novel DELM-AE consists of one input layer, two random hidden mapping layers for encoding feature, and one output layer for decoding feature. When stacking DELM-AE in the hierarchical structure, they can construct an H-DELM model, where the input of the current AE is the feature representation learned by the previous one, but the output is identical to the original input information and is not the input. Hence, the H-DELM can reproduce the original input data as much as possible to learn more expressive and compact feature. They validate their method on various widely public datasets, and the results demonstrate that H-DELM can bring significant performance improvements in terms of classification accuracy and robustness compared with existing relevant multilayer ELM and other deep learning algorithms at a slight computational cost.

关键词： learning (artificial intelligence) pattern classification random hidden mapping layers output layer decoding feature stacking DELM-AE hierarchical structure H-DELM model feature representation original input information original input data expressive feature compact feature deep learning algorithms extreme learning machine multilayer network architecture lead extremely fast training speed expressive representation learning method hierarchical ELM H-ELM novel architectural unit double random hidden layers ELM auto-encoder novel DELM-AE research developments relevant multilayer ELM

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：