Publications by authors named "Yun Fu"

176 Publications

Cladding softened fiber for sensitivity enhancement of distributed acoustic sensing.

Opt Express 2021 Mar;29(6):8216-8222

Fiber-optic distributed acoustic sensing (DAS) technology with high spatial and strain resolutions has been widely used in many practical applications. New methods to enhance the phase sensitivity of sensing fiber are worth exploring to further improve DAS performances, although the standard single-mode fiber (SSMF) has been widely used for DAS technology. In this work, we propose and demonstrate the concept of enhancing the phase sensitivity of DAS by softening the cladding of the sensing fiber, for the first time. The theoretical analysis indicates that softening sensing fiber cladding is an effective way to improve phase sensitivity. Thus, we fabricated cladding softened fibers (CSFs) and tested their phase sensitivities experimentally. According to the results, it is found that the phase sensitivity of the CSF with 0.48 WT% phosphorus-doping concentration and 80 µm cladding diameter is 22% and 54% higher than that of the non-phosphorus-doping fiber with 80 µm cladding diameter and SSMF, respectively. The results show that by reducing fiber cladding Young's modulus with higher phosphorus-doping concentration, the DAS phase sensitivity can be enhanced effectively, verifying the theoretical analysis. Also, we found that the phase sensitivity enhancement of the sensing fiber has a linear relationship with the cladding phosphorus-doping concentration, i.e. Young's modulus. In conclusion, the reported CSF paves a way for improving the DAS phase sensitivity and would be applied to other major optical fiber sensing systems as a better sensing element over SSMF due to the enhancement in the elasto-optical effect of the sensing fiber.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1364/OE.417360DOI Listing
March 2021

The Roles of Immunoregulatory Networks in Severe Drug Hypersensitivity.

Front Immunol 2021 26;12:597761. Epub 2021 Feb 26.

College of Medicine, Chang Gung University, Taoyuan, Taiwan.

The immunomodulatory effects of regulatory T cells (Tregs) and co-signaling receptors have gained much attention, as they help balance immunogenic and immunotolerant responses that may be disrupted in autoimmune and infectious diseases. Drug hypersensitivity has a myriad of manifestations, which ranges from the mild maculopapular exanthema to the severe Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), and drug reaction with eosinophilia and systemic symptoms/drug-induced hypersensitivity syndrome (DRESS/DIHS). While studies have identified high-risk human leukocyte antigen (HLA) allotypes, the presence of the HLA allotype at risk is not sufficient to elicit drug hypersensitivity. Recent studies have suggested that insufficient regulation by Tregs may play a role in severe hypersensitivity reactions. Furthermore, immune checkpoint inhibitors, such as anti-CTLA-4 or anti-PD-1, in cancer treatment also induce hypersensitivity reactions including SJS/TEN and DRESS/DIHS. Taken together, mechanisms involving both Tregs as well as coinhibitory and costimulatory receptors may be crucial in the pathogenesis of drug hypersensitivity. In this review, we summarize the currently implicated roles of co-signaling receptors and Tregs in delayed-type drug hypersensitivity in the hope of identifying potential pharmacologic targets.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3389/fimmu.2021.597761DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7953830PMC
February 2021

Survey on the Analysis and Modeling of Visual Kinship: A Decade in the Making.

IEEE Trans Pattern Anal Mach Intell 2021 Mar 2;PP. Epub 2021 Mar 2.

Kinship recognition is a challenging problem with many practical applications. With much progress and milestones having been reached after ten years - we are now able to survey the research and create new milestones. We review the public resources and data challenges that enabled and inspired many to hone-in on the views of automatic kinship recognition in the visual domain. The different tasks are described in technical terms and syntax consistent across the problem domain and the practical value of each discussed and measured. State-of-the-art methods for visual kinship recognition problems, whether to discriminate between or generate from, are examined. As part of such, we review systems proposed as part of a recent data challenge held in conjunction with the 2020 IEEE Conference on Automatic Face and Gesture Recognition. We establish a stronghold for the state of progress for the different problems in a consistent manner. This survey will serve as the central resource for the work of the next decade to build upon. For the tenth anniversary, the demo code is provided for the various kin-based tasks. Detecting relatives with visual recognition and classifying the relationship is an area with high potential for impact in research and practice.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TPAMI.2021.3063078DOI Listing
March 2021

Large-Scale Subspace Clustering by Independent Distributed and Parallel Coding.

IEEE Trans Cybern 2021 Feb 26;PP. Epub 2021 Feb 26.

Subspace clustering is a popular method to discover underlying low-dimensional structures of high-dimensional multimedia data (e.g., images, videos, and texts). In this article, we consider a large-scale subspace clustering (LS²C) problem, that is, partitioning million data points with a millon dimensions. To address this, we explore an independent distributed and parallel framework by dividing big data/variable matrices and regularization by both columns and rows. Specifically, LS²C is independently decomposed into many subproblems by distributing those matrices into different machines by columns since the regularization of the code matrix is equal to a sum of that of its submatrices (e.g., square-of-Frobenius/ℓ₁-norm). Consensus optimization is designed to solve these subproblems in a parallel way for saving communication costs. Moreover, we provide theoretical guarantees that LS²C can recover consensus subspace representations of high-dimensional data points under broad conditions. Compared with the state-of-the-art LS²C methods, our approach achieves better clustering results in public datasets, including a million images and videos.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TCYB.2021.3052056DOI Listing
February 2021

Deep Learning for Fashion Style Generation.

IEEE Trans Neural Netw Learn Syst 2021 Feb 26;PP. Epub 2021 Feb 26.

In this article, we work on generating fashion style images with deep neural network algorithms. Given a garment image, and single or multiple style images (e.g., flower, blue and white porcelain), it is a challenge to generate a synthesized clothing image with single or mix-and-match styles due to the need to preserve global clothing contents with coverable styles, to achieve high fidelity of local details, and to conform different styles with specific areas. To address this challenge, we propose a fashion style generator (FashionG) framework for the single-style generation and a spatially constrained FashionG (SC-FashionG) framework for mix-and-match style generation. Both FashionG and SC-FashionG are end-to-end feedforward neural networks that consist of a generator for image transformation and a discriminator for preserving content and style globally and locally. Specifically, a global-based loss is calculated based on full images, which can preserve the global clothing form and design. A patch-based loss is calculated based on image patches, which can preserve detailed local style patterns. We develop an alternating patch-global optimization methodology to minimize these losses. Compared with FashionG, SC-FashionG employs an additional spatial constraint to ensure that each style is blended only onto a specific area of the clothing image. Extensive experiments demonstrate the effectiveness of both single-style and mix-and-match style generations.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TNNLS.2021.3057892DOI Listing
February 2021

Adenosquamous carcinoma of the breast: a population-based study.

Breast Cancer 2021 Feb 12. Epub 2021 Feb 12.

Department of Breast Surgery, Shanghai First Maternity and Infant Hospital, Tongji University School of Medicine, Shanghai, 200040, China.

Background: To summarize the clinicopathological characteristics, prognosis, and management of breast adenosquamous carcinoma (ASC).

Methods: A population-based study was performed using retrospectively extracted data from the Surveillance, Epidemiology, and End Results database for breast cancer patients with histological diagnoses of ASC, infiltrating duct carcinoma (IDC) and squamous cell carcinoma (SCC) from 2004 to 2016.

Results: ASC presented similar tumor size but low histological grade and less lymph node metastasis compared to IDC. ASC expressed less positive rate of hormone receptors and barely HER2, which was similar with SCC. ASC patients underwent the similar surgical and systematic treatment as IDC, only with less radiotherapy. Median follow-up data of 78 months showed that the prognosis of IDC patients was better than that of ASC patients (all p < 0.05 for BCSM and OS). ASC was not an independent prognosis factor of breast cancer. After propensity score matching (PSM), no significant difference in BCSM nor OS was observed between ASC and IDC groups. In HR-negative patients, the prognosis of ASC was similar with that of IDC, and both were superior to SCC. In HR-positive patients, the 5-year survival rate of ASC was 63.5%, which was far less than that in ASC of HR-negative (81.0%). Multivariate analysis showed that older age (age > 60) and advanced AJCC-stage were independent factors of poor prognosis in ASC, breast-conserving surgery was also ideally suited for ASC.

Conclusions: ASC has unique clinicopathological characteristics and prognosis. It is imperative to focus on a more precise and personalized treatment management of ASC patients.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s12282-021-01222-3DOI Listing
February 2021

Correction to: Mir-20a-5p induced WTX deficiency promotes gastric cancer progressions through regulating PI3K/AKT signaling pathway.

J Exp Clin Cancer Res 2021 Jan 28;40(1):46. Epub 2021 Jan 28.

Department of Pathology, Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences, Guangzhou, 510000, Guangdong Province, People's Republic of China.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s13046-020-01809-2DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7841984PMC
January 2021

Geometrically Editable Face Image Translation With Adversarial Networks.

IEEE Trans Image Process 2021 12;30:2771-2783. Epub 2021 Feb 12.

Recently, image-to-image translation has received increasing attention, which aims to map images in one domain to another specific one. Existing methods mainly solve this task via a deep generative model that they focus on exploring the bi-directional or multi-directional relationship between specific domains. Those domains are often categorized by attribute-level or class-level labels, which do not incorporate any geometric information in learning process. As a result, existing methods are incapable of editing geometric contents during translation. They also neglect to utilize higher-level and instance-specific information to further guide the training process, leading to a great deal of unrealistic synthesized images of low fidelity, especially for face images. To address these challenges, we formulate the general image translation problem as multi-domain mappings in both geometric and attribute directions within an image set that shares a same latent vector. Particularly, we propose a novel Geometrically Editable Generative Adversarial Networks (GEGAN) model to solve this problem for face images by leveraging facial semantic segmentation to explicitly guide its geometric editing. In details, input face images are encoded to their latent representations via a variational autoencoder, a segmentor network is designed to impose semantic information on the generated images, and multi-scale regional discriminators are employed to force the generator to pay attention to the details of key components. We provide both quantitative and qualitative evaluations on CelebA dataset to demonstrate our ability of the geometric modification and our improvement in image fidelity.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TIP.2021.3052084DOI Listing
February 2021

Characterizations of Ischemic Stroke Complications in Cardiac Myxoma Patients at a Single Institution in Eastern China.

Neuropsychiatr Dis Treat 2021 7;17:33-40. Epub 2021 Jan 7.

Department of Neurology, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou, Zhejiang Province 325000, People's Republic of China.

Background: Cardiac myxoma is the most common primary cardiac tumor. Even though it rarely causes a stroke, it is an important risk factor. Here, we compared our clinical experience in managing myxoma patients who developed stroke complications with those who did not present with this condition at the First Affiliated Hospital of Wenzhou Medical University.

Patients And Methods: The medical records were reviewed of 160 cardiac myxoma patients who were treated in our facility from January 2006 to December 2019. They were separated into either a stroke group or non-stroke group.

Results: Cardiac obstructive symptoms, embolic events and constitutional symptoms were observed in 92 (57.7%), 25 (15.6%) and 18 (11.2%) patients, respectively. Among 23 cardiac myxoma ischemic stroke patients, hypoesthesia (60.9%), hemiparesis (56.5%) and facial paresis (47.8%) were the three most common neurological symptoms. The middle cerebral artery (82.6%) was the most commonly affected vessel, whereas 73.9% of the ischemic patients had multiple stroke lesions. Binary logistic regression analysis showed that coronary heart disease and tumor sizes were independently associated in the stroke group (p <0.05). The 10 years cumulative survival rate was 87.9% for all patients after surgical intervention. There was no significant difference in the 10 years cumulative survival rate between the two groups (80.0% vs 88.9%, p =0.274 > 0.05).

Conclusion: The three most common neurological symptoms (hypoesthesia, hemiparesis and facial paresis), the middle cerebral artery and multiple lesions involvements were the definitive markers of patients afflicted with cardiac myxoma stroke. Small tumor sizes were independently associated with these patients. Surgical resection is a relatively safe procedure for treating both the stroke and non-stroke patients.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.2147/NDT.S280641DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7800702PMC
January 2021

Generative Partial Multi-View Clustering With Adaptive Fusion and Cycle Consistency.

IEEE Trans Image Process 2021 14;30:1771-1783. Epub 2021 Jan 14.

Nowadays, with the rapid development of data collection sources and feature extraction methods, multi-view data are getting easy to obtain and have received increasing research attention in recent years, among which, multi-view clustering (MVC) forms a mainstream research direction and is widely used in data analysis. However, existing MVC methods mainly assume that each sample appears in all the views, without considering the incomplete view case due to data corruption, sensor failure, equipment malfunction, etc. In this study, we design and build a generative partial multi-view clustering model with adaptive fusion and cycle consistency, named as GP-MVC, to solve the incomplete multi-view problem by explicitly generating the data of missing views. The main idea of GP-MVC lies in two-fold. First, multi-view encoder networks are trained to learn common low-dimensional representations, followed by a clustering layer to capture the shared cluster structure across multiple views. Second, view-specific generative adversarial networks with multi-view cycle consistency are developed to generate the missing data of one view conditioning on the shared representation given by other views. These two steps could be promoted mutually, where the learned common representation facilitates data imputation and the generated data could further explores the view consistency. Moreover, an weighted adaptive fusion scheme is implemented to exploit the complementary information among different views. Experimental results on four benchmark datasets are provided to show the effectiveness of the proposed GP-MVC over the state-of-the-art methods.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TIP.2020.3048626DOI Listing
January 2021

Resveratrol induces human colorectal cancer cell apoptosis by activating the mitochondrial pathway via increasing reactive oxygen species.

Mol Med Rep 2021 03 5;23(3). Epub 2021 Jan 5.

Department of Pathology, School of Basic Medical Sciences, Southern Medical University, Guangzhou, Guangdong 510515, P.R. China.

Colorectal cancer (CRC) is the third most commonly diagnosed malignancy and the second leading cause of cancer‑related mortality worldwide according to Global Cancer Statistics 2018. Resveratrol (RSV) is a phenolic compound that possesses anticancer functions against various types of cancer, including breast and gastric cancer. However, the functions and mechanism underlying RSV in CRC are not completely understood. The present study aimed to investigate the anticancer effects and mechanism underlying RSV in CRC cells by conducting Cell Counting Kit‑8, apoptosis, reactive oxygen species (ROS) and western blotting assays. The results suggested that RSV dose‑dependently inhibited CRC cell viability, and increased cell apoptosis and ROS levels compared with the control group. The protein expression levels of Bax, cytochrome c, cleaved caspase‑9 and cleaved caspase‑3 were upregulated, whereas Bcl‑2 expression levels were downregulated in RSV‑treated CRC cells compared with control cells. The results indicated that RSV might activate the mitochondrial apoptotic pathway by increasing ROS release. The present study suggested that RSV possessed antitumour activity against CRC by modulating an ROS‑mediated mitochondrial apoptotic pathway.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3892/mmr.2020.11809DOI Listing
March 2021

Representative Task Self-Selection for Flexible Clustered Lifelong Learning.

IEEE Trans Neural Netw Learn Syst 2020 Dec 21;PP. Epub 2020 Dec 21.

Consider the lifelong machine learning paradigm whose objective is to learn a sequence of tasks depending on previous experiences, e.g., knowledge library or deep network weights. However, the knowledge libraries or deep networks for most recent lifelong learning models are of prescribed size and can degenerate the performance for both learned tasks and coming ones when facing with a new task environment (cluster). To address this challenge, we propose a novel incremental clustered lifelong learning framework with two knowledge libraries: feature learning library and model knowledge library, called Flexible Clustered Lifelong Learning (FCL³). Specifically, the feature learning library modeled by an autoencoder architecture maintains a set of representation common across all the observed tasks, and the model knowledge library can be self-selected by identifying and adding new representative models (clusters). When a new task arrives, our FCL³ model firstly transfers knowledge from these libraries to encode the new task, i.e., effectively and selectively soft-assigning this new task to multiple representative models over feature learning library. Then: 1) the new task with a higher outlier probability will be judged as a new representative, and used to redefine both feature learning library and representative models over time; or 2) the new task with lower outlier probability will only refine the feature learning library. For model optimization, we cast this lifelong learning problem as an alternating direction minimization problem as a new task comes. Finally, we evaluate the proposed framework by analyzing several multitask data sets, and the experimental results demonstrate that our FCL³ model can achieve better performance than most lifelong learning frameworks, even batch clustered multitask learning models.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TNNLS.2020.3042500DOI Listing
December 2020

Learnable Subspace Clustering.

IEEE Trans Neural Netw Learn Syst 2020 Dec 11;PP. Epub 2020 Dec 11.

This article studies the large-scale subspace clustering (LS²C) problem with millions of data points. Many popular subspace clustering methods cannot directly handle the LS²C problem although they have been considered to be state-of-the-art methods for small-scale data points. A simple reason is that these methods often choose all data points as a large dictionary to build huge coding models, which results in high time and space complexity. In this article, we develop a learnable subspace clustering paradigm to efficiently solve the LS²C problem. The key concept is to learn a parametric function to partition the high-dimensional subspaces into their underlying low-dimensional subspaces instead of the computationally demanding classical coding models. Moreover, we propose a unified, robust, predictive coding machine (RPCM) to learn the parametric function, which can be solved by an alternating minimization algorithm. Besides, we provide a bounded contraction analysis of the parametric function. To the best of our knowledge, this article is the first work to efficiently cluster millions of data points among the subspace clustering methods. Experiments on million-scale data sets verify that our paradigm outperforms the related state-of-the-art methods in both efficiency and effectiveness.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TNNLS.2020.3040379DOI Listing
December 2020

Design, synthesis, and biological activity of novel 1,2,4-oxadiazole derivatives.

BMC Chem 2020 Nov 22;14(1):68. Epub 2020 Nov 22.

State Key Laboratory Breeding Base of Green Pesticide and Agricultural Bioengineering, Key Laboratory of Green Pesticide and Agricultural Bioengineering, Ministry of Education, Guizhou University, Guiyang, 550025, China.

Background: Plant diseases seriously threaten food security, it is urgent to discover efficient and low-risk chemical pesticides. 1,2,4-Oxadiazole derivatives exhibit broad spectrum of agricultural biological activities. For discovering novel molecules with excellent agricultural activities, novel 1,2,4-oxadiazole derivatives were synthesized and evaluated for their agricultural activities.

Result: Bioassays results showed that the title compounds exhibited moderate nematocidal activity against Meloidogyne incognita and anti-fungal activity to Rhizoctonia solani. It's worth noting that compounds 5m, 5r, 5u, 5v, 5x and 5y showed strong antibacterial effects on Xanthomonas oryzae pv. oryzae (Xoo), with EC values of 36.25, 24.14, 28.82, 19.44, 25.37 and 28.52 μg/mL, respectively, superior to bismerthiazol (BMT, EC = 77.46 μg/mL) and thiodiazole copper (TDC, EC = 99.31 μg/mL). Compounds 5p, 5u and 5v exhibited excellent antibacterial ability against Xanthomonas oryzae pv. oryzicola (Xoc), with EC values of 31.40, 19.04 and 21.78 μg/mL, respectively, better than that of BMT (EC = 68.50 μg/mL) and TDC (EC = 91.05 μg/mL). In addition, compound 5v exerted moderate antibacterial effects on rice bacterial leaf blight.

Conclusions: Twenty-six novel 1,2,4-oxadiazole derivatives were obtained and their biological activities were evaluated. Compound 5u and 5v exhibited excellent antibacterial activity Xoo and Xoc. These results indicated that 1,2,4-oxadiazole derivatives containing a trifluoromethyl pyridine moiety could be as potential alternative templates for discovering novel antibacterial agents.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s13065-020-00722-1DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7680602PMC
November 2020

Auricular therapy improves gastrointestinal function in patients with gynecological laparoscopic surgery: A protocol for systematic review and meta analysis.

Medicine (Baltimore) 2020 Dec;99(49):e23421

Sandun District Zhejiang Hospital, Hangzhou, Zhejiang Province, China.

Background: Gynecological laparoscopic surgery is the main method to treat gynecological diseases, but postoperative gastrointestinal reactions are more common in patients. Auricular therapy, as a characteristic therapy of Traditional Chinese Medicine, can improve gastrointestinal symptoms such as nausea and vomiting by stimulating the conduction of acupoints through the nervous system on internal organs, but there are studies questioning the efficacy of auricular therapy. Therefore, the purpose of this study is to prove the efficacy and safety of auricular therapy in promoting gastrointestinal function recovery after gynecological laparoscopic surgery, and to provide reference value for future clinical practice.

Methods: To search English databases (PubMed, Excerpta Medical Database [Embase], Web of Science, the Cochrane Library) and Chinese databases (Chinese National Knowledge Internet [CNKI], WanFang, Viper, Chinese Biomedical Literature Database) by computer, and conduct a randomized controlled trial on the effect of aural point therapy on gastrointestinal function recovery of patients after gynecological laparoscopic surgery from the establishment of the database to October 2020. Two researchers independently evaluate the quality of the included studies and extract the data, and meta-analysis of the included literature is carried out using RevMan5.3 software.

Results: In this study, the efficacy and safety of auricular therapy in the recovery of gastrointestinal function after gynecological laparoscopic surgery are evaluated from the aspects of first anal exhaust time, bowel sound recovery time, and incidence of gastrointestinal complications.

Conclusion: This study will provide reliable evidence-based evidence for auricular therapy in the treatment of gastrointestinal function after gynecologic laparoscopic surgery.

Ethics And Dissemination: Private information from individuals will not be published. This systematic review also does not involve endangering participant rights. Ethical approval was not required. The results may be published in a peer-reviewed journal or disseminated at relevant conferences.

Osf Registration Number: DOI 10.17605 / OSF.IO / ZSPGA.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1097/MD.0000000000023421DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7717805PMC
December 2020

Vehicle and Person Re-Identification With Support Neighbor Loss.

IEEE Trans Neural Netw Learn Syst 2020 Oct 23;PP. Epub 2020 Oct 23.

One of the key tasks for an intelligent visual surveillance system is to automatically re-identify objects of interest, e.g., persons or vehicles, from nonoverlapping camera views. This demand incurs the vast investigation of person re-identification (re-ID) and vehicle re-ID techniques, especially those deep learning-based ones. While most recent algorithms focus on designing new convolutional neural networks, less attention is paid to the loss functions, which are of vital roles as well. Triplet loss and softmax loss are the two losses that are extensively used, both of which, however, have limitations. Triplet loss optimizes the model to produce features with which samples from the same class have higher similarity than those from different classes. The problem of triplet loss is that the number of triplets to be constructed grows cubically with training samples, which causes scalability issue, unstable performance, and slow convergence. Softmax loss has favorable scalable property and is widely used for large-scale classification problems. However, since Softmax loss only aims to separate well training classes, its performance for re-ID tasks is not desirable because the model is tested to measure the similarity of samples from unseen classes. We propose the support neighbor (SN) loss, which avoids the limitations of the abovementioned two losses. Unlike triplet loss that is calculated based on triplets, SN loss is derived from K-nearest neighbors (SNs) of anchor samples. The SNs of an anchor are unique, containing more valuable contextual information and neighborhood structure of the anchor, and thus contribute to more stable performance and reliable embedding from image space to feature space. Based on the SNs, a softmax-like separation term and a squeeze term are proposed, which encourage interclass separation and intraclass compactness, respectively. Experiments show that SN loss surpasses triplet and softmax losses with the same backbone network and reaches the state-of-the-art performance for both person and vehicle re-ID using a ResNet50 backbone when combined with training tricks.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TNNLS.2020.3029299DOI Listing
October 2020

Detecting Lesional Granulysin Levels for Rapid Diagnosis of Cytotoxic T lymphocyte-Mediated Bullous Skin Disorders.

J Allergy Clin Immunol Pract 2021 Mar 8;9(3):1327-1337.e3. Epub 2020 Oct 8.

Department of Dermatology, Drug Hypersensitivity Clinical and Research Center, Chang Gung Memorial Hospital, Linkou, Taipei, Keelung, Taoyuan, Taiwan; Cancer Vaccine and Immune Cell Therapy Core Laboratory, Chang Gung Memorial Hospital, Linkou, Taoyuan, Taiwan; Chang Gung Immunology Consortium, Chang Gung Memorial Hospital, Linkou, Taoyuan, Taiwan; Graduate Institute of Clinical Medical Sciences, College of Medicine, Chang Gung University, Kwei-Shan, Taoyuan, Taiwan; College of Medicine, Chang Gung University, Taoyuan, Taiwan; Department of Dermatology, Xiamen Chang Gung Hospital, Xiamen, China; Chang Gung Immunology Consortium, Chang Gung Memorial Hospital and Chang Gung University, Taoyuan, Taiwan; Whole-Genome Research Core Laboratory of Human Diseases, Chang Gung Memorial Hospital, Keelung, Taiwan; Immune-Oncology Center of Excellence, Chang Gung Memorial Hospital, Linkou, Taiwan; Department of Dermatology, Beijing Tsinghua Chang Gung Hospital, School of Clinical Medicine, Tsinghua University, Beijing, China; Department of Dermatology, Ruijin Hospital, School of Medicine, Shanghai Jiaotong University, Shanghai, China; Genomic Medicine Core Laboratory, Chang Gung Memorial Hospital, Linkou, Taiwan. Electronic address:

Background: Bullous skin disorders are induced by different pathomechanisms and several are emergent, including Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN). Rapid diagnostic methods for SJS/TEN or cytotoxic T-lymphocyte (CTL)-mediated bullous disorders are crucial for early treatment. Granulysin, primarily expressed by CTLs, is a specific cytotoxic protein responsible for SJS/TEN and similar skin reactions.

Objective: To assess granulysin levels in blister fluids to differentiate SJS/TEN and similar CTL-mediated bullous reactions from other autoimmune bullous disorders.

Methods: Using ELISA, we measured granulysin in blister fluids from patients with bullous skin disorders, including SJS/TEN, erythema multiforme major, bullous fixed-drug eruption, bullous lupus erythematosus, paraneoplastic pemphigus, pemphigus vulgaris, bullous pemphigoid, purpura fulminans-related bullae, and hand-foot syndrome/hand-foot-skin reactions. We compared serum and blister granulysin levels in patients with SJS/TEN presenting varying severity, monitoring serial granulysin levels from acute to late stages.

Results: Overall, 144 patients presenting with bullous skin disorders were enrolled. Blister granulysin levels (mean ± SD) in CTL-mediated disorders, including TEN (n = 28; 3938.7 ± 3475.7), SJS-TEN overlapping (n = 22; 1440.4 ± 1179.6), SJS (n = 14; 542.0 ± 503.2), erythema multiforme major (n = 7; 766.3 ± 1073.7), generalized bullous fixed-drug eruption (n = 10; 720.4 ± 858.3), and localized bullous fixed-drug eruption (n = 16; 69.0 ± 56.4), were significantly higher than in non-CTL-mediated bullous disorders (P < .0001), including bullous lupus erythematosus (n = 3; 22.7 ± 20.1), paraneoplastic pemphigus (n = 3; 20.3 ± 8.6), pemphigus vulgaris (n = 3; 4.4 ± 2.8), bullous pemphigoid (n = 18; 4.0 ± 2.7), purpura fulminans (n = 4; 5.9 ± 5.5), and hand-foot syndrome/hand-foot-skin reactions (n = 6; 4.6 ± 3.5). Blister granulysin levels correlated with clinical severity of SJS/TEN (P < .0001).

Conclusions: Determination of blister granulysin levels is a noninvasive and useful tool for rapid differential diagnosis of SJS/TEN and other similar CTL-mediated bullous skin disorders for treatment selection.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jaip.2020.09.048DOI Listing
March 2021

Mir-20a-5p induced WTX deficiency promotes gastric cancer progressions through regulating PI3K/AKT signaling pathway.

J Exp Clin Cancer Res 2020 Oct 8;39(1):212. Epub 2020 Oct 8.

1Department of Pathology, Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences, Guangzhou, 510000, Guangdong Province, People's Republic of China.

Background: The X-linked gene WTX (also called AMER1) has been reported to function as a tumour suppressor gene in Wilms' tumour. In our previous study, WTX expression was shown to be significantly reduced in gastric cancer (GC), but the function and mechanism associated with WTX loss had yet to be fully elucidated.

Methods: WTX expression and clinical significance were father analyzed in GC and control normal gastric tissues, and validated in public databases. The candidate pathway which was regulated by WTX during GC progression was searched by KEGG pathway analysis. The miRNA which monitored WTX expression was screened by miRNA microarray. After verified the pathway and miRNA both in vitro and in vivo, the relationship of miRNA, WTX and the downstream pathway were analyzed by Western blot, immunohistochemistry, RT-PCR, Co-immunoprecipitation (Co-IP), and luciferase analyses.

Results: The results showed that WTX serves as a tumour suppressor gene in GC. The loss of WTX which is associated with the aggressiveness of GC by promoting GC cell proliferation in vitro and high metastasis in vivo. Furthermore, WTX expression was positively correlated with the overall survival of GC patients. Microarray assays, bioinformatics analysis, and verification experiments showed that WTX loss activates the PI3K/AKT/mTOR pathway and promotes GC cell proliferation and invasion. And the aberrant miR-20a-5p upregulation contributes to WTX loss in GC, which stimulates PI3K phosphorylation to activate PI3K/AKT/mTOR signaling pathway and promoted GC progression.

Conclusions: The results of the present study elucidated the mechanism of GC progression, which is at least partially caused by aberrant miR-20a-5p upregulation leading to the inhibition of WTX expression and PI3K/AKT/mTOR signaling pathway activation. These findings provide a comprehensive understanding of the action of the miR-20a-5p/WTX/PI3K/AKT/mTOR signaling pathway in the progression and metastasis of GC.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s13046-020-01718-4DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7545863PMC
October 2020

Association of Psoriasis with Colorectal Cancer: A Systematic Review and Meta-Analysis.

J Am Acad Dermatol 2020 Sep 28. Epub 2020 Sep 28.

Department of Dermatology, Chang Gung Memorial Hospital, Linkou, Taoyuan, Taiwan; College of Medicine, Chang Gung University, Taoyuan, Taiwan. Electronic address:

Background: The relation between psoriasis and colorectal cancer (CRC) was largely unclear.

Objective: To investigate the association of psoriasis with CRC.

Methods: A systematic review and meta-analysis of observational studies that examined the association of psoriasis with CRC was performed. We searched MEDLINE and Embase on March 24, 2020 for relevant studies. The Newcastle-Ottawa scale was used to evaluate the risk of bias of included studies. We conducted random-effects model meta-analysis and subgroup analysis on different gender.

Results: We included nine cohort studies with 10,544,609 subjects. We found a significantly increased risk for CRC in patients with psoriasis (hazard ratio (HR) 1.16, 95% confidence interval (CI) 1.08-1.24). Subgroups analysis based on gender found significantly increased risk for CRC in female psoriatic patients (HR 1.41, 95% CI 1.16-1.72) but not in male ones (HR 1.18, 95% CI 0.92-1.50).

Limitation: No data on psoriatic arthritis.

Conclusions: Psoriatic patients have an increased risk for CRC. Gastroenterology consultation and colonoscopic examination is indicated for psoriatic patients presenting with bowel symptoms.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jaad.2020.09.050DOI Listing
September 2020

Effects of Omega-3 Polyunsaturated Fatty Acid Supplementation on Non-Alcoholic Fatty Liver: A Systematic Review and Meta-Analysis.

Nutrients 2020 Sep 11;12(9). Epub 2020 Sep 11.

Department of Dermatology, Chang Gung Memorial Hospital, Linkou, Taoyuan 33305, Taiwan.

(1) Aim: Non-alcoholic fatty liver disease (NAFLD) is a prevalent disease worldwide. Omega-3 polyunsaturated fatty acids (n-3 PUFAs) bear anti-inflammatory action and can ameliorate hyperlipidemia. We wish to appraise the effects of n-3 PUFAs supplement on NAFLD. (2) Methods: We searched CENTRAL, Embase, and MEDLINE on 29 March 2020 for randomized control trials (RCTs) on the effects of n-3 PUFAs supplementation in treating NAFLD. The Cochrane Collaboration's tool was used to assess the risk of bias of included RCTs. (3) Results: We included 22 RCTs with 1366 participants. The risk of bias of included RCTs was generally low or unclear. n-3 PUFAs supplementation significantly reduced liver fat compared with placebo (pooled risk ratio 1.52; 95% confidence interval (CI) 1.09 to 2.13). n-3 PUFAs supplementation also significantly improved the levels of triglyceride, total cholesterol, high-density lipoprotein, and body-mass index, with pooled mean difference and 95% CI being -28.57 (-40.81 to -16.33), -7.82 (-14.86 to -0.79), 3.55 (1.38 to 5.73), and -0.46 (-0.84 to -0.08), respectively. (4) Conclusions: The current evidence supports the effects of n-3 PUFAs supplementation in improving fatty liver. n-3 PUFAs supplementation may also improve blood lipid levels and obesity.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3390/nu12092769DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7551292PMC
September 2020

Self-accelerating HO-responsive Plasmonic Nanovesicles for Synergistic Chemo/starving therapy of Tumors.

Theranostics 2020 9;10(19):8691-8704. Epub 2020 Jul 9.

State Key Laboratory of Molecular Engineering of Polymers, Department of Macromolecular Science, Fudan University, Shanghai 200438, China.

Nanoscale vehicles responsive to abnormal variation in tumor environment are promising for use in targeted delivery of therapeutic drugs specifically to tumor sites. Herein, we report the design and fabrication of self-accelerating HO-responsive plasmonic gold nanovesicles (GVs) encapsulated with tirapazamine (TPZ) and glucose oxidase (GOx) for synergistic chemo/starving therapy of cancers. Gold nanoparticles were modified with HO-responsive amphiphilic block copolymer PEG--PABE by ligand exchange. The TPZ and GOx loaded GVs (TG-GVs) were prepared through the self-assembly of PEG--PABE -grafted nanoparticles together with TPZ and GOx by solvent displacement method. In response to HO in tumor, the TG-GVs dissociate to release the payloads that are, otherwise, retained inside the vesicles for days without noticeable leakage. The released GOx enzymes catalyze the oxidation of glucose by oxygen in the tumor tissue to enhance the degree of hypoxia that subsequently triggers the reduction of hypoxia-activated pro-drug TPZ into highly toxic free radicals. The HO generated in the GOx-catalyzed reaction also accelerate the dissociation of vesicles and hence the release rate of the cargoes in tumors. The drug-loaded GVs exhibit superior tumor inhibition efficacy in 4T1 tumor-bearing mice owing to the synergistic effect of chemo/starvation therapy, in addition to their use as contrast agents for computed tomography imaging of tumors. This nanoplatform may find application in managing tumors deeply trapped in viscera or other important tissues that are not compatible with external stimulus (e.g. light).
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.7150/thno.45392DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7392001PMC
July 2020

Uric acid fluctuation had no effect on renal function among gout patients.

Nutr Metab Cardiovasc Dis 2020 09 20;30(10):1833-1839. Epub 2020 Jun 20.

Department of Public Health, and Department of Anesthesiology, the Second Affiliated Hospital of Zhejiang University School of Medicine, Hangzhou, China. Electronic address:

Background And Aims: To investigate the effects of serum uric acid (SUA) level and its fluctuation on renal dysfunction in gout patients.

Methods And Results: Data on gout patients was collected from Huzhou city electronic medical record system data sharing platform, and information about relevant diagnoses, prescriptions, biochemical indexes and imaging characteristics was extracted. The gout patients with baseline normal renal function were enrolled in this analysis, and the estimated glomerular filtration rate (eGFR) < 60 mL/min/1.73 m was defined as renal dysfunction. The generalized estimating equation and Cox regression analysis were used. A total of 1009 patients with gout were enrolled. Compared with the reference group (normal baseline SUA with endpoint SUA to be < 6 mg/dL), endpoint SUA ≥ 10 mg/dL was associated with an increased risk of renal dysfunction (baseline normal SUA group: HR [95% CI] = 3.28 [1.21, 8.91]; baseline high SUA group: HR [95% CI] = 3.01 [1.43, 6.35]). Subgroup analysis of 771 SUA stable gout patients demonstrated that SUA levels at 8-10 (excluding 10), and ≥10 mg/dL were significantly associated with an increased risk for renal dysfunction, with HR [95%CI] to be 1.99 [1.05, 3.77], and 2.98 [1.38, 6.43], respectively.

Conclusion: Regardless of the baseline SUA level, SUA >10 mg/dL was a significant risk factor for renal dysfunction. SUA between 6 and 10 mg/dL was a potential risk factor for renal dysfunction. No significant correlation of SUA fluctuation and renal function was found.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.numecd.2020.06.008DOI Listing
September 2020

Connexin 43 Modulates the Cellular Resistance to Paclitaxel via Targeting β-Tubulin in Triple-Negative Breast Cancer.

Onco Targets Ther 2020 10;13:5323-5335. Epub 2020 Jun 10.

Department of Breast Surgery, Shanghai First Maternity and Infant Hospital, Tongji University School of Medicine, Shanghai, People's Republic of China.

Background: Triple-negative breast cancer has become an intricate part and hotspot in the clinical and experimental research. Connexins, serving as functional proteins in gap junctions, play an important role in tumorigenesis, cell proliferation and metastasis.

Methods: We constructed and employed the Connexin 43 (Cx43) overexpression lentiviral vectors and Cx43 siRNA in paclitaxel-treated MDA-MB-231 cells. We performed the experiments of clonal formation and flow cytometry to gauge the effect of paclitaxel on cellular behaviors and immunofluorescence and subsequent quantitative RT-PCR and Western blot to examine the expression of genes and corresponding proteins. Experiments of scrape loading/dye transfer were utilized to explore the gap junctions. The targets of Cx43 were identified via the experiments of co-immunoprecipitation (Co-IP), GST pull-down assays and proximal ligation assay (PLA).

Results: The results showed that Cx43 hindered cell proliferation and promoted apoptosis in the paclitaxel-treated MDA-MB-231 cells. Overexpressed Cx43 suppressed the expression of resistance genes such as BRCP, Txr-1, α-tubulin and β-tubulin and promoted the expression of apoptosis gene as TSP-1 and Bcl-2. Cx43 was also positively related to ITGα9 and negatively related to ITGαV and ITGα11. The gap junctions altered magnificently under different expressions of Cx43, which indicated that Cx43 could promote the number of intercellular gap junctions. The immunofluorescent experiment revealed that both of Cx43 and β-tubulin were mainly localized in the cytoplasm. The assays of Co-IP and GST pull-down demonstrated that there existed a direct interaction between Cx43 and β-tubulin. Furthermore, the result of PLA also showed that Cx43 interacts with β-tubulin in MDA-MB-231 cells.

Conclusion: Overexpression of Cx43 could modulate the cellular resistance to paclitaxel via targeting β-tubulin in triple-negative breast cancer.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.2147/OTT.S229076DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7294565PMC
June 2020

Molecular epidemiology and changes in genotype diversity of norovirus infections in acute gastroenteritis patients in Huzhou, China, 2018.

J Med Virol 2020 Jun 30. Epub 2020 Jun 30.

Microbiological Laboratory, Huzhou Center for Disease Control and Prevention, Huzhou, China.

Norovirus is an important causative agent of acute gastroenteritis worldwide, affecting people of all ages. Stool samples collected from patients with clinical symptoms of acute gastroenteritis in all age groups at the diarrhea outpatient department of the First People's Hospital in Huzhou were analyzed to gain insight into the prevalence and genetic characteristics of norovirus. From January to December 2018, a total of 551 specimens were screened for norovirus by real-time reverse transcription-polymerase chain reaction (RT-PCR). RT-PCR was used for genomic amplification and sequencing of the RNA-dependent RNA polymerase and capsid gene of the positive samples. Genotypes of norovirus were assigned using the norovirus Noronet typing tool and phylogenetic analysis. About 100 (18.1%) specimens were identified as norovirus positive. GII genogroup was the main genogroup identified (83.0%; 83/100). About 42 (42.0%) samples were successfully sequenced and genotyped by RT-PCR. Since one of the samples was dual infection, so we got 43 virus finally. Nine norovirus GII genotypes and four norovirus GI genotypes were detected in Huzhou during our research period. The main two norovirus GII genotypes were GII.2[P16] (54.8%; 23/43) and GII.17[P17] (11.9%; 5/43). We characterized the molecular epidemiology of norovirus infection in acute gastroenteritis patients during 2018. GII genogroup was the main genogroup identified. The dominance norovirus genotype circulating in the population of Huzhou was GII.2[P16] in 2018.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/jmv.26247DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7692952PMC
June 2020

Machine learning-aided quantification of antibody-based cancer immunotherapy by natural killer cells in microfluidic droplets.

Lab Chip 2020 06;20(13):2317-2327

Department of Pharmaceutical Sciences, Northeastern University, 360 Huntington Avenue, Boston, MA, USA.

Natural killer (NK) cells have emerged as an effective alternative option to T cell-based immunotherapies, particularly against liquid (hematologic) tumors. However, the effectiveness of NK cell therapy has been less than optimal for solid tumors, partly due to the heterogeneity in target interaction leading to variable anti-tumor cytotoxicity. This paper describes a microfluidic droplet-based cytotoxicity assay for quantitative comparison of immunotherapeutic NK-92 cell interaction with various types of target cells. Machine learning algorithms were developed to assess the dynamics of individual effector-target cell pair conjugation and target death in droplets in a semi-automated manner. Our results showed that while short contacts were sufficient to induce potent killing of hematological cancer cells, long-lasting stable conjugation with NK-92 cells was unable to kill HER2+ solid tumor cells (SKOV3, SKBR3) significantly. NK-92 cells that were engineered to express FcγRIII (CD16) mediated antibody-dependent cellular cytotoxicity (ADCC) selectively against HER2+ cells upon addition of Herceptin (trastuzumab). The requirement of CD16, Herceptin and specific pre-incubation temperature served as three inputs to generate a molecular logic function with HER2+ cell death as the output. Mass proteomic analysis of the two effector cell lines suggested differential changes in adhesion, exocytosis, metabolism, transport and activation of upstream regulators and cytotoxicity mediators, which can be utilized to regulate specific functionalities of NK-92 cells in future. These results suggest that this semi-automated single cell assay can reveal the variability and functional potency of NK cells and may be used to optimize immunotherapeutic efficacy for preclinical analyses.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1039/d0lc00158aDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7938931PMC
June 2020

Bidirectional Association Between Psoriasis and Obstructive Sleep Apnea: A Systematic Review and Meta-Analysis.

Sci Rep 2020 04 3;10(1):5931. Epub 2020 Apr 3.

Department of Dermatology, Chang Gung Memorial Hospital, Linkou, Taoyuan, Taiwan.

The link between psoriasis and obstructive sleep apnea (OSA) has not been confirmed. We aimed to investigate the relationship between psoriasis and OSA. We conducted a systematic review and meta-analysis of case-control, cross-sectional, and cohort studies on the association between psoriasis and OSA. We searched MEDLINE and Embase for relevant studies on May 11, 2019. The Newcastle-Ottawa Scale was used to evaluate the risk of bias of included studies. We performed random-effects model meta-analysis to calculate pooled odds ratio (ORs) with 95% confidence intervals (CIs) for case-control and cross-sectional studies as well as pooled incidence rate ratio (IRR) with 95% CIs for cohort studies in association between psoriasis and OSA. A total of 4 case-control or cross-sectional studies and 3 cohort studies with a total of 5,840,495 subjects were included. We identified a significantly increased odds for OSA in psoriasis patients (pooled OR 2.60; 95% CI 1.07-6.32), and significantly increased risk for psoriasis in OSA patients (pooled IRR 2.52; 95% CI 1.89-3.36). In conclusion, our study identified a bidirectional association between psoriasis and OSA. Sleep quality should be inquired in patients with psoriasis. Respirologist consultation or polysomnography may be indicated for those presenting with night snoring, recurrent awaking, and excessive daytime sleepiness.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41598-020-62834-xDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7125081PMC
April 2020

Integrated principal component analysis denoising technique for phase-sensitive optical time domain reflectometry vibration detection.

Appl Opt 2020 Jan;59(3):669-675

This paper presents an integrated principal component analysis (IPCA) technique for denoising phase-sensitive optical time domain reflectometry (-OTDR) sensing data for vibration detection. As one of the key distributed optical fiber sensing technologies, it has attracted great attention, mainly due to its high sensitivity, fast response time, dynamic range, and vibration detection abilities. To enhance vibration detection along the sensing fiber, an appropriate denoising method must be carefully selected. Hence, the PCA that can effectively reduce noise on signals while preserving significant details of the denoised signal is identified. It was then applied on the said signal after digital down-conversion where the noise was greatly reduced. Then angle and phase unwrapping was performed and the vibration was clearly detected with a significant enhancement of the signal-to-noise ratio. As proof of concept, the theoretical analysis and an experimental demonstration of a vibration sensing range of 800 m are presented.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1364/AO.59.000669DOI Listing
January 2020

Continual Multiview Task Learning via Deep Matrix Factorization.

IEEE Trans Neural Netw Learn Syst 2021 Jan 4;32(1):139-150. Epub 2021 Jan 4.

The state-of-the-art multitask multiview (MTMV) learning tackles a scenario where multiple tasks are related to each other via multiple shared feature views. However, in many real-world scenarios where a sequence of the multiview task comes, the higher storage requirement and computational cost of retraining previous tasks with MTMV models have presented a formidable challenge for this lifelong learning scenario. To address this challenge, in this article, we propose a new continual multiview task learning model that integrates deep matrix factorization and sparse subspace learning in a unified framework, which is termed deep continual multiview task learning (DCMvTL). More specifically, as a new multiview task arrives, DCMvTL first adopts a deep matrix factorization technique to capture hidden and hierarchical representations for this new coming multiview task while accumulating the fresh multiview knowledge in a layerwise manner. Then, a sparse subspace learning model is employed for the extracted factors at each layer and further reveals cross-view correlations via a self-expressive constraint. For model optimization, we derive a general multiview learning formulation when a new multiview task comes and apply an alternating minimization strategy to achieve lifelong learning. Extensive experiments on benchmark data sets demonstrate the effectiveness of our proposed DCMvTL model compared with the existing state-of-the-art MTMV and lifelong multiview task learning models.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TNNLS.2020.2977497DOI Listing
January 2021

Reverse Attention Based Residual Network for Salient Object Detection.

IEEE Trans Image Process 2020 Jan 22. Epub 2020 Jan 22.

Benefiting from the quick development of deep convolutional neural networks, especially fully convolutional neural networks (FCNs), remarkable progresses have been achieved on salient object detection recently. Nevertheless, these FCNs based methods are still challenging to generate high resolution saliency maps, and also not applicable for subsequent applications due to their heavy model weights. In this paper, we propose a compact and efficient deep network with high accuracy for salient object detection. Firstly, we propose two strategies for initial prediction, one is a new designed multi-scale context module, the other is incorporating hand-crafted saliency priors. Secondly, we employ residual learning to refine it progressively by only learning the residual in each side-output, which can be achieved with few convolutional parameters, therefore leads to high compactness and high efficiency. Finally, we further design a novel reverse attention block to guide side-output residual learning in a top-down manner. Specifically, the current predicted salient regions are erased from each side-output feature, thus the missing object parts and details can be efficiently learned from these unerased regions, which results in high resolution and accuracy. Extensive experimental results on seven benchmark datasets demonstrate that the proposed network performs favorably against the state-of-the-art methods, and with advantages in terms of simplicity, efficiency and model size.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TIP.2020.2965989DOI Listing
January 2020

Residual Dense Network for Image Restoration.

IEEE Trans Pattern Anal Mach Intell 2020 Jan 23;PP. Epub 2020 Jan 23.

Recently, deep convolutional neural network (CNN) has achieved great success for image restoration (IR) and provided hierarchical features at the same time. However, most deep CNN based IR models do not make full use of the hierarchical features from the original low-quality images, thereby resulting in relatively-low performance. In this work, we propose a novel and efficient residual dense network (RDN) to address this problem in IR, by making a better tradeoff between efficiency and effectiveness in exploiting the hierarchical features from all the convolutional layers. Specifically, we propose residual dense block (RDB) to extract abundant local features via densely connected convolutional layers. RDB further allows direct connections from the state of preceding RDB to all the layers of current RDB, leading to a contiguous memory mechanism. To adaptively learn more effective features from preceding and current local features and stabilize the training of wider network, we proposed local feature fusion in RDB. After fully obtaining dense local features, we use global feature fusion to jointly and adaptively learn global hierarchical features in a holistic way. We demonstrate the effectiveness of RDN with several representative IR applications, single image super-resolution, Gaussian image denoising, image compression artifact reduction, and image deblurring. Experiments on benchmark and real-world datasets show that our RDN achieves favorable performance against state-of-the-art methods for each IR task quantitatively and visually.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TPAMI.2020.2968521DOI Listing
January 2020