检索结果-内蒙古大学图书馆

Adversarial-Metric Learning for Audio-Visual Cross-Modal Matching

IEEE TRANSACTIONS ON MULTIMEDIA 2022年 24卷 338-351页

作者： Zheng, Aihua Hu, Menglan Jiang, Bo Huang, Yan Yan, Yan Luo, Bin Minist Educ Key Lab Intelligent Comp & Signal Proc Anhui Prov Key Lab Multimodal Cognit Computat Hefei Peoples R China Anhui Univ Sch Comp Sci & Technol Hefei 230601 Peoples R China Anhui Univ Inst Phys Sci & Informat Technol Hefei 230601 Peoples R China Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China IIT Dept Comp Sci Chicago IL 60616 USA

Audio-visual matching aims to learn the intrinsic correspondence between image and audio clip. Existing works mainly concentrate on learning discriminative features, while ignore the cross-modal heterogeneous issue between audio and visual modalities. To deal with this issue, we propose a novel Adversarial-Metric Learning (AML) model for audio-visual matching. AML aims to generate a modality-independent representation for each person in each modality via adversarial learning, while simultaneously learns a robust similarity measure for cross-modality matching via metric learning. By integrating the discriminative modality-independent representation and robust cross-modality metric learning into an end-to-end trainable deep network, AML can overcome the heterogeneous issue with promising performance for audio-visual matching. Experiments on the various audio-visual learning tasks, including audio-visual matching, audio-visual verification and audio-visual retrieval on benchmark dataset demonstrate the effectiveness of the proposed AML model. The implementation codes are available on https://***/MLanHu/AML.

关键词： Visualization Task analysis Measurement Speech recognition Videos Location awareness Image recognition Adversarial learning audio-visual matching cross-modal learning metric learning

来源：评论

学校读者我要写书评

暂无评论

On the Equivalence of Linear Discriminant Analysis and Least Squares Regression

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2024年第4期35卷 5710-5720页

作者： Nie, Feiping Chen, Hong Xiang, Shiming Zhang, Changshui Yan, Shuicheng Li, Xuelong Northwestern Polytech Univ Sch Comp Sci Sch Artificial Intelligence Opt & Elect iOPEN Xian 710072 Shaanxi Peoples R China Northwestern Polytech Univ Key Lab Intelligent Interact & Applicat Minist Ind & Informat Technol Xian 710072 Shaanxi Peoples R China Chinese Acad Sci Inst Automat Dept Natl Lab Pattern Recognit Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 100049 Peoples R China Tsinghua Univ THUAI Inst Artificial Intelligence Beijing 100084 Peoples R China Tsinghua Univ Beijing Natl Res Ctr Informat Sci & Technol BNRist Dept Automat Beijing 100084 Peoples R China Natl Univ Singapore Sea AI Lab SAIL Singapore 117583 Singapore

Studying the relationship between linear discriminant analysis (LDA) and least squares regression (LSR) is of great theoretical and practical significance. It is well-known that the two-class LDA is equivalent to an LSR problem, and directly casting multiclass LDA as an LSR problem, however, becomes more challenging. Recent study reveals that the equivalence between multiclass LDA and LSR can be established based on a special class indicator matrix, but under a mild condition which may not hold under the scenarios with low-dimensional or oversampled data. In this article, we show that the equivalence between multiclass LDA and LSR can be established based on arbitrary linearly independent class indicator vectors and without any condition. In addition, we show that LDA is also equivalent to a constrained LSR based on the data-dependent indicator vectors. It can be concluded that under exactly the same mild condition, such two regressions are both equivalent to the null space LDA method. Illuminated by the equivalence of LDA and LSR, we propose a direct LDA classifier to replace the conventional framework of LDA plus extra classifier. Extensive experiments well validate the above theoretic analysis.

关键词： Eigenvalues and eigenfunctions principal component analysis Null space Training data Symmetric matrices Sun Optimization Least squares regression (LSR) linear discriminant analysis (LDA) linear regression minimum distance classifier (MDC) null space LDA

来源：评论

学校读者我要写书评

暂无评论

Functional connectome gradient of prefrontal cortex as biomarkers of high risk for internet gaming disorder

引用

NEUROIMAGE 2025年 306卷 121010-121010页

作者： Wen, Xinwen Yue, Lirong Du, Zhe Zhao, Jiahao Ge, Mengjiao Yuan, Cunfeng Wang, Hongmei He, Qinghua Yuan, Kai Xidian Univ Sch Life Sci & Technol Xian 710071 Shaanxi Peoples R China Southwest Univ Fac Psychol MOE Key Lab Cognit & Personal Chongqing 400715 Peoples R China Drug Rehabil Adm Minist Justice Beijing Peoples R China Inner Mongolia Peoples Hosp Dept Med Imaging Hohhot 010017 Peoples R China Inner Mongolia Univ Sci & Technol Sch Informat Engn Inner Mongolia Key Lab Pattern Recognit & Intellig Baotou 014010 Inner Mongolia Peoples R China Minist Educ Engn Res Ctr Mol & Neuro Imaging Xian 710071 Shaanxi Peoples R China Xidian Univ Sch Life Sci & Technol Xian Key Lab Intelligent Sensing & Regulat Transsc Xian 710126 Shaanxi Peoples R China

Adolescents and young adults are considered a high-risk group for internet gaming disorder (IGD). Early screening for high-risk individuals with IGD and exploring the underlying neural mechanisms is an effective strategy to reduce the harm of IGD. We recruited 219 non-internet gaming addicted college students and evaluated them with magnetic resonance imaging, followed by a two-year longitudinal follow-up. We used functional connectome gradient (FCG) to capture the macroscopic hierarchical organization of human brain. Canonical correlation analysis was employed to identify components mapping relationships between FCG and behavioral scores. Consequently, K-means clustering was used to define distinct subtypes. The risk of developing IGD and FCG patterns were compared among the subtypes. Three subtypes were identified and subtype 3 exhibited the highest risk for developing IGD according to the occurrence rates of IGD two years later: (1) subtype 1 (5.3 %, 4 participants), (2) subtype 2 (10.8 %, 9 participants), (3) subtype 3 (20 %, 12 participants). The abnormal FCG in the inferior frontal gyrus and posterior cingulate cortex at baseline were observed in subtype 3, which were correlated with impulsivity. These findings advanced understanding of the biological and behavioral heterogeneity associated with developing of IGD, and represented a promising step toward the prediction of high-risk individuals.

关键词： Internet gaming disorder Adolescents and young adults prefrontal cortex Impulsivity Risk factor Functional connectome gradient Magnetic resonance imaging

来源：评论

学校读者我要写书评

暂无评论

Research on Fast recognition of Vulnerable Traffic participants in intelligent Connected Vehicles on Edge Computing

引用

JOURNAL OF CIRCUITS SYSTEMS AND COMpUTERS 2023年第3期32卷 2350046-2350046页

作者： Gu, Musong Lyu, Jingjing Li, Zhongwen Yan, Zihan Fan, Wenjie Chengdu Univ Coll Comp Sci Chengdu Peoples R China Chengdu Univ Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat P Chengdu Peoples R China Southern Univ Sci & Technol Dept Math Shenzhen Peoples R China

Real-time and fast recognition of all kinds of traffic participants in intelligent driving has always been a major difficulty in the research of internet of vehicles. With the advent of edge computing, we try to deploy an image recognition algorithm directly to the intelligent vehicles. However, the original image recognition algorithm is difficult to be directly deployed on the vehicles due to limited edge device resources. Based on this, a fast recognition model of vulnerable traffic participants based on depthwise separable convolutional neural network (DSCYOLO) is proposed in this paper. The algorithm can significantly reduce the convolutional parameter quantity and computing load, making it suitable for deployment on the vehicle-mounted edge embedded devices. In order to validate the effectiveness of the proposed method, its simulation results are compared with the main target detection models Faster R-CNN, SSD and YOLOv3. The results show that the recognition time of the proposed model is reduced by 80.28%, 66.80% and 86.74%, respectively, on the basis of a relatively high recognition precision. The model can realize real-time detection and fast recognition of vulnerable traffic participants, so as to avoid a large number of traffic accidents. It has significant social and economic benefits.

关键词： intelligent connected vehicle YOLOv3 edge computing real-time traffic warning

来源：评论

学校读者我要写书评

暂无评论

Optical solitons and single traveling wave solutions for the Triki-Biswas equation describing monomode optical fibers

引用

OpTIK 2022年 258卷 168835-168835页

作者： Li, Zhao Lian, Zhigang Chengdu Univ Coll Comp Sci Chengdu 610106 Peoples R China Chengdu Univ Key Lab Pattern Recognit & Intelligent Informat Chengdu 610106 Peoples R China

In this paper, the polynomial complete discriminant system method is introduced to construct optical solitons and single traveling wave solutions for the Triki-Biswas equation, which is one of the most important mathematical and physical equation usually used to simulate phenomena in nonlinear optics, quantum mechanics and plasma physics. The Jacobi elliptic function solutions, dark solitons solutions, bright solitons solutions, singular solitons solutions and other solutions have been presented.

关键词： Triki-Biswas equation Optical solitons Traveling wave solution Complete discrimination system

来源：评论

学校读者我要写书评

暂无评论

Vector-Mediated Cancer Gene Therapy Reduces Toxicity and Inhibition of Lung Carcinoma Growth in Nude Mice

引用

AppLIED BIOCHEMISTRY AND BIOTECHNOLOGY 2024年第1期196卷 261-274页

作者： Khalaf, Ahmad Taha Wan, Jun Wei, Hu Fubing, Shen Zainol, Jamaludin Kadir, Samiah Yasmin Abdul Liang, Min Chengdu Univ Basic Med Coll Chengdu 610106 Sichuan Peoples R China Affiliated Hosp Chengdu 610081 Peoples R China Chengdu Univ Clin Med Coll Chengdu 610081 Peoples R China Chengdu Univ Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat Pr Chengdu 610106 Sichuan Peoples R China Hosp Wuhan Univ Wuhan Hubei Peoples R China Widad Univ Coll Darul Makmur Pahang Malaysia Sichuan Univ West China Hosp 4 Dept Thorac Surg Chengdu 610041 Peoples R China

Replication-competent oncolytic adenovirus (TOA2) gene therapy is a recently introduced anti-tumor treatment regimen with superior results. The biodistribution studies of virus vector-based medicine seem more cautious and have been given much attention recently in terms of its quality and safety in preclinical trials. The current study determined the biodistribution and safety of a replication-competent adenovirus in different organs to predict its toxicity threshold. The present study has used TOA2, while biodistribution analysis was performed in human lung carcinoma A549-induced tumor-bearing nude mice model. Intratumoral injection was applied onto tumor-bearing mice with the adenovirus (3x10(10) Vp per mouse). Mice were sacrificed at the end of the experiment and the organs were dissected. Biodistribution analysis was done with complete hexon gene detection in each organ using quantitative real-time polymerase chain reaction (qRT-pCR). The biodistribution and concentration profiles showed that the TOA2 is well distributed in the entire tumor tissue. After dose 3 at day 11, the concentration of the virus has increased in the tumor tissue from 2240.54 (+/- 01.69) copies/100 ng genome to 13,120.28 (+/- 88.21) copies/100 ng genome on the 18th day, which eventually approached 336.45 (+/- 23.41) copies/100ng genome on the day 36. On the contrary, the concentration of the same decreased in the order of the liver, kidney, spleen, lung, and heart over time but no distributional traces in gonads. But the concentration found decreased dramatically in blood and other organs, while at the end of the experiment no detectable distribution was seen besides tumor tissue. The study confirms that adenovirus-based tumor therapy using conditionally replicating competent oncolytic TOA2 exhibited great efficiency with no toxicity at all.

关键词： Oncolytic virus TOA2 Biodistribution Lung carcinoma A549 cell line

来源：评论

学校读者我要写书评

暂无评论

Exploiting compressed sensing and polynomial-based progressive secret image sharing for visually secure image selection encryption with authentication

引用

JOURNAL OF KING SAUD UNIVERSITY-COMpUTER AND informatION SCIENCES 2022年第10期34卷 9252-9272页

作者： Gan, Zhihua Song, Shiping Zhou, Lin Han, Daojun Fu, Jiangyu Chai, Xiuli Henan Univ Inst Intelligent Network Syst Sch Software Intelligent Data Proc Engn Res Ctr Henan Prov Kaifeng 475004 Peoples R China Henan Key Lab Cyberspace Situat Awareness Zhengzhou 450001 Peoples R China Henan Univ Sch Artificial Intelligence Henan Key Lab Big Data Anal & Proc Zhengzhou 450046 Peoples R China Henan Univ Inst Image Proc & Pattern Recognit Sch Comp & Informat Engn Kaifeng 475004 Peoples R China

Image encryption is an effective method to protect private images by converting them into meaningless ones, but many current image encryption algorithms still have security and efficiency problems. This paper combines compressed sensing (CS) and secret image sharing (SIS) to propose a certifiable visually secure image selection encryption. Firstly, a selective encryption based on multi-task convolutional neural network (SE-MTCNN) is presented to distinguish and encrypt the sensitive and non-sensitive information of plain images. Subsequently, the SIS algorithm is used to decompose the resulting plain image into multiple shadow images to disperse the risk of cryptographic image loss and resist hacking attacks. Herein, a pseudo-convolutional sliding scrambling based on chaotic sequences driven by plaintext (pCSS-CSDp) is proposed to eliminate the correlation between adjacent image pixels. Finally, the IWTM-embedding algorithm is used to embed shadow images into carrier images to get multiple visually meaningful cipher images. Additionally, a new authentication scheme based on bit-level cyclic shifts and dynamic combination (AS-BCSDC) is introduced, and the obtained authentication information is fused into cipher images to prevent the deception of illegal users and malicious tampering of cloud platform. Experiments demonstrate the effectiveness of our algorithm, and it may be applied for image secure communication.

关键词： Compressed sensing (CS) Secret image sharing (SIS) Integer wavelet transform (IWT)-M-embedding Image encryption

来源：评论

学校读者我要写书评

暂无评论

Debiased Video-Text Retrieval via Soft positive Sample Calibration

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2023年第9期33卷 5257-5270页

作者： Zhang, Huaiwen Yang, Yang Qi, Fan Qian, Shengsheng Xu, Changsheng Inner Mongolia Univ Coll Comp Sci Hohhot 010021 Peoples R China Natl & Local Joint Engn Res Ctr Intelligent Inform Hohhot 010021 Peoples R China Inner Mongolia Key Lab Mongolian Informat Proc Tec Hohhot 010021 Peoples R China Tianjin Univ Technol Sch Comp Sci & Engn Tianjin 300384 Peoples R China Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 100049 Peoples R China

With the emergence of enormous videos on various video apps, semantic video-text retrieval has become a critical task for improving the user experience. The primary paradigm for video-text retrieval learns the semantic video-text representations in a common space by pulling the positive samples close to the query and pushing the negative samples away. However, in practice, the video-text datasets contain only the annotations of positive samples. The negative samples are randomly drawn from the entire dataset. There may exist soft positive samples, which are sampled as negatives but share the same semantics as positive samples. Indiscriminately enforcing the model to push all the negative samples away from the query leads to inaccurate supervision and then misleads the video-text feature representation learning. In this paper, we introduce debiased video-text retrieval objectives that calibrate the punishment of soft positive samples. In particular, we propose a novel uncertainty measure framework to estimate the credibility of negative samples for each instance. Then, the reliability of negative samples is used to find the soft positive samples and rescale their contribution within video-text retrieval losses, including triplet loss and contrastive loss. Experimental results on five widely used datasets demonstrate that our debiased video-text retrieval objectives achieve significant performance improvements and establish a new state-of-the-art.

关键词： debias soft positive samples

来源：评论

学校读者我要写书评

暂无评论

Robust Video-Text Retrieval Via Noisy pair Calibration

引用

IEEE TRANSACTIONS ON MULTIMEDIA 2023年 25卷 8632-8645页

作者： Zhang, Huaiwen Yang, Yang Qi, Fan Qian, Shengsheng Xu, Changsheng Inner Mongolia Univ Coll Comp Sci Mongolia 010031 Peoples R China Natl & Local Joint Engn Res Ctr Intelligent Inform Mongolia 010031 Peoples R China Inner Mongolia Key Lab Mongolian Informat Proc Tec Hohhot 010021 Peoples R China Tianjin Univ Technol Sch Comp Sci & Engn Tianjin 300384 Peoples R China Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 100049 Peoples R China

Video-text retrieval is a fundamental task in managing the emerging massive amounts of video data. The main challenge focuses on learning a common representation space for videos and queries where the similarity measurement can reflect the semantic closeness. However, existing video-text retrieval models may suffer from the following noise in the common space learning procedure: First, the video-text correspondences in positive pairs may not be exact matches. The crowdsourcing annotation for existing datasets leads to inevitable tagging noise for non-expert annotators. Second, the learning of video-text representation is based on the negative samples randomly sampled. Instances that are semantically similar to the query may be incorrectly categorized as negative samples. To alleviate the adverse impact of these noisy pairs, we propose a novel robust video-text retrieval method that protects the model from noisy positive and negative pairs by identifying and calibrating noisy pairs with their uncertainty score. In particular, we propose a noisy pair identifier, which divides the training dataset into noisy and clean subsets based on the estimated uncertainty of each pair. Then, with the help of uncertainties, we calibrate the two types of noisy pairs with an adaptive margin triplet loss and a weighted triplet loss function, respectively. To verify the effectiveness of our methods, we conduct extensive experiments on three widely used datasets. Experimental results show that the proposed robust video-text retrieval methods successfully identify and calibrate the noisy pairs and improve retrieval performance.

关键词： Noise calibration uncertainty video text retrieval

来源：评论

学校读者我要写书评

暂无评论

Optimal Design of Thermal Cycling Reliability for plastic Ball Grid Array Assembly Via Finite Element Method and Taguchi Method

引用

JOURNAL OF ELECTRONIC pACKAGING 2024年第1期146卷 011009-011009页

作者： Gao, Chao Huang, Chunyue Ying, Liang Guilin Univ Elect Technol Sch Mech & Elect Engn 1 Jinji Rd Guilin 541004 Peoples R China Chengdu Univ Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat Pr 2025 Chengluo Ave Chengdu 610106 Peoples R China

A finite element simulation analysis model was developed for a plastic ball grid array (pBGA) assembly to analyze its behavior under thermal cyclic loading conditions. The stress distribution in the SAC305 solder joints at different locations within the array was investigated by using ANAND constitutive equations. Subsequently, the thermal fatigue life of the key solder joints was quantified. The study also examined the influence of the solder joint diameter, substrate thickness, solder joint height, printed circuit board (pCB) thickness, and mold compound height on solder joint stress. Optimization of assembly parameters was achieved through the application of the Taguchi method. An extensive analysis was conducted using different assembly parameter combinations, employing the L9(34) orthogonal array design to explore the thermal cycling effects. The computed average Von Mises stress Delta sigma for the critical thin-layer elements of solder joints located in hazardous positions within the assembly was notably affected by variations in the solder joint height, substrate thickness, solder joint diameter, and mold compound height. This impact ranked in descending order of significance as solder joint height, substrate thickness, solder joint diameter, and mold compound height. The optimal parameter combination determined was a solder joint height of 0.70 mm, a solder joint diameter of 0.85 mm, a substrate thickness of 0.51 mm, and a mold compound height of 1.12 mm. Implementing this optimized configuration led to a significant 4.07% reduction in average stress Delta sigma for the critical thin-layer elements within hazardous solder joints. Moreover, the extension of the thermal fatigue life was notably improved, achieving an impressive 51.72% increase.

关键词： pBGA assembly ANAND thermal cyclic Taguchi method optimization

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：