4,901 Hits in 2.5 sec

Recurrent Regression for Face Recognition [article]

Yang Li, Wenming Zheng, Zhen Cui
2016 arXiv   pre-print
To address the sequential changes of images including poses, in this paper we propose a recurrent regression neural network(RRNN) framework to unify two classic tasks of cross-pose face recognition on still images and video-based face recognition. To imitate the changes of images, we explicitly construct the potential dependencies of sequential images so as to regularize the final learning model. By performing progressive transforms for sequentially adjacent images, RRNN can adaptively memorize
more » ... and forget the information that benefits for the final classification. For face recognition of still images, given any one image with any one pose, we recurrently predict the images with its sequential poses to expect to capture some useful information of others poses. For video-based face recognition, the recurrent regression takes one entire sequence rather than one image as its input. We verify RRNN in static face dataset MultiPIE and face video dataset YouTube Celebrities(YTC). The comprehensive experimental results demonstrate the effectiveness of the proposed RRNN method.
arXiv:1607.06999v1 fatcat:4gu4qjudkfeytg2upgjkzbkvau

A deep learning method for solving high-order nonlinear soliton equation [article]

Shikun Cui, Zhen Wang, Jiaqi Han, Xinyu Cui
2021 arXiv   pre-print
We propose effective scheme of deep learning method for high-order nonlinear soliton equation and compare the activation function for high-order soliton equation. The neural network approximates the solution of the equation under the conditions of differential operator, initial condition and boundary condition. We apply this method to high-order nonlinear soliton equation, and verify its efficiency by solving the fourth-order Boussinesq equation and the fifth-order Korteweg de Vries equation.
more » ... e results show that deep learning method can solve the high-order nonlinear soliton equation and reveal the interaction between solitons.
arXiv:2106.11024v1 fatcat:7xfvbk7pmzf47dsqz6scrffdou

Association between ERα gene Pvu II polymorphism and breast cancer susceptibility

Zhen-lian Zhang, Cui-zhen Zhang, Yan Li, Zhen-hui Zhao, Shun-e Yang
2018 Medicine  
Estrogen has played an important role in the development of breast cancer. ER-a PvuII gene polymorphism is in close association with the occurrence risk of breast cancer, but no consensus has been achieved currently. Methods: PubMed, Embase, China National Knowledge Infrastructure (CNKI) database, Wanfang database, and VIP database were retrieved to collect the case-control studies on association between ERa gene Pvu II polymorphism and breast cancer risk published before September 1, 2017.
more » ... astle-Ottawa Scale (NOS) was used to assess the quality of the literatures, Stata 14.0 software was applied for meta-analysis, and the pooled odds ratio (OR) and 95% confidence interval (95% CI) were calculated. The subgroup analysis was performed to assess the confounding factors, followed by assessment of publication bias and sensitivity analysis. Results: A total of 26 studies were enrolled in the analysis based on inclusion criteria, which included 15,360 patients and 26,423 controls. The results demonstrated that ERa gene Pvu II polymorphism was in significant association with the decrease of breast cancer risk in 3 genetic models (C vs T, OR = 0.962, 95% CI = 0.933-0.992, P = .012; CC vs TT, OR = 0.911, 95% CI = 0.856-0.969, P = .003; CC vs TT/CT, OR = 0.923, 95% CI = 0.874-0.975, P = .004). Subgroup analysis was conducted on the basis of ethnicity and source of controls, whose results illustrated that ERa gene Pvu II polymorphism was in significant association with the decrease of breast cancer risk in Asians rather than in Caucasians (CC vs TT, OR = 0.862, 95% CI = 0.750-0.922, P = .038; CC vs TT/CT, OR = 0.851, 95% CI = 0.755-0.959, P = .008). In population-based subgroup rather than in hospital-based subgroup, ERa gene Pvu II polymorphism was in significant association with the decrease of breast cancer risk in the allele model, homozygous model, dominant model, and recessive model (C vs T, OR = 0.943, 95% CI = 0.911-0.977, P = .001; CC vs TT, OR = 0.878, 95% CI = 0.817-0.944, P = .000; CC/CT vs TT, OR = 0.936, 95% CI = 0.881-0.994, P = .031; CC vs TT/CT, OR = 0.902, 95% CI = 0.847-0.960, P = .001). Conclusion: ERa gene Pvu II polymorphism exerts an important function in the progression of breast cancer. Abbreviations: CBM = China Biomedicine, CI = confidence interval, CNKI = China National Knowledge Infrastructure, ER = estrogen receptors, ERa = estrogen receptor a, ERE = estrogen response element, HB = hospital-based, HWE = Hardy-Weinberg equilibrium, MALDI-TOF = matrix-assisted laser desorption ionization time-of-flight, NOS = Newcastle-Ottawa Scale, OR = odds ratio, PB = population-based study, PB = population-based, PCR-RFLP = polymerase chain reaction-restriction fragment length polymorphism, SNPs = single nucleotide polymorphisms.
doi:10.1097/md.0000000000010317 pmid:29702977 pmcid:PMC5944501 fatcat:wvie3ffecrgezcsvazrudqfkqq

Spatial Transformer Point Convolution [article]

Yuan Fang, Chunyan Xu, Zhen Cui, Yuan Zong, Jian Yang
2020 arXiv   pre-print
Point clouds are unstructured and unordered in the embedded 3D space. In order to produce consistent responses under different permutation layouts, most existing methods aggregate local spatial points through maximum or summation operation. But such an aggregation essentially belongs to the isotropic filtering on all operated points therein, which tends to lose the information of geometric structures. In this paper, we propose a spatial transformer point convolution (STPC) method to achieve
more » ... otropic convolution filtering on point clouds. To capture and represent implicit geometric structures, we specifically introduce spatial direction dictionary to learn those latent geometric components. To better encode unordered neighbor points, we design sparse deformer to transform them into the canonical ordered dictionary space by using direction dictionary learning. In the transformed space, the standard image-like convolution can be leveraged to generate anisotropic filtering, which is more robust to express those finer variances of local regions. Dictionary learning and encoding processes are encapsulated into a network module and jointly learnt in an end-to-end manner. Extensive experiments on several public datasets (including S3DIS, Semantic3D, SemanticKITTI) demonstrate the effectiveness of our proposed method in point clouds semantic segmentation task.
arXiv:2009.01427v1 fatcat:vohmdyqtebfmpbvr6uyawh4aea

Gaussian-Induced Convolution for Graphs [article]

Jiatao Jiang, Zhen Cui, Chunyan Xu, Jian Yang
2018 arXiv   pre-print
Niepert, Ahmed, and Kutzkov 2016; Song et al. 2018 ) and social network (Gomez, Chiem, and Delvenne 2017; Orsini, Baracchi, and Frasconi 2017) to images/videos (Marino, Salakhutdinov, and Gupta 2016; Cui  ... 
arXiv:1811.04393v1 fatcat:bx43rawmqfa6lnjysijr2hjgau

mir-500-Mediated GAD67 Downregulation Contributes to Neuropathic Pain

Zhen-Zhen Huang, Jia-You Wei, Han-Dong Ou-Yang, Dai Li, Ting Xu, Shao-Ling Wu, Xiao-Long Zhang, Cui-Cui Liu, Chao Ma, Wen-Jun Xin
2016 Journal of Neuroscience  
Neuropathic pain is a common neurobiological disease involving multifaceted maladaptations ranging from gene modulation to synaptic dysfunction, but the interactions between synaptic dysfunction and the genes that are involved in persistent pain remain elusive. In the present study, we found that neuropathic pain induced by the chemotherapeutic drug paclitaxel or L5 ventral root transection significantly impaired the function of GABAergic synapses of spinal dorsal horn neurons via the reduction
more » ... of the GAD67 expression. We also found that mir-500 expression was significantly increased and involved in the modulation of GAD67 expression via targeting the specific site of Gad1 gene in the dorsal horn. In addition, knock-out of mir-500 or using mir-500 antagomir rescued the GABAergic synapses in the spinal dorsal horn neurons and attenuated the sensitized pain behavior in the rats with neuropathic pain. To our knowledge, this is the first study to investigate the function significance and the underlying molecular mechanisms of mir-500 in the process of neuropathic pain, which sheds light on the development of novel therapeutic options for neuropathic pain. Neuropathic pain is a common neurobiological disease involving multifaceted maladaptations ranging from gene modulation to synaptic dysfunction, but the underlying molecular mechanisms remain elusive. The present study illustrates for the first time a mir-500-mediated mechanism underlying spinal GABAergic dysfunction and sensitized pain behavior in neuropathic pain induced by the chemotherapeutic drug paclitaxel or L5 ventral root transection, which sheds light on the development of novel therapeutic options for neuropathic pain.
doi:10.1523/jneurosci.0646-16.2016 pmid:27277808 fatcat:sa5hinwtzzgjfahuudpy7zekwa

Spoof Plasmon Hybridization [article]

Jingjing Zhang, Zhen Liao, Yu Luo, Xiaopeng Shen, Stefan A. Maier, Tie Jun Cui
2016 arXiv   pre-print
Supporting Information Enormous Field Enhancement in Hybrid Spoof Plasmonic System Jingjing Zhang, Zhen Liao, Yu Luo *, Xiaopeng Shen, Stefan A.  ...  Maier *, and Tie Jun Cui * * To whom correspondence should be addressed; E-mail:,, We consider the corrugated structures with spiral grooves.  ... 
arXiv:1610.03645v1 fatcat:46dz3dj6ufhqtpyyjtm7f3uesi

On sums of subsets of Chen primes [article]

Zhen Cui, Hongze Li, Boqing Xue
2012 arXiv   pre-print
In this paper we show that if A is a subset of Chen primes with positive relative density α, then A+A must have positive upper density at least cα e^-c^'(1/α)^2/3((1/α))^1/3 in the natural numbers.
arXiv:1206.2471v1 fatcat:rf2kjrwmznbivgbekuynxdgpem

Multi-pretrained Deep Neural Network [article]

Zhen Hu, Zhuyin Xue, Tong Cui, Shiqiang Zong, Chenglong He
2016 arXiv   pre-print
Pretraining is widely used in deep neutral network and one of the most famous pretraining models is Deep Belief Network (DBN). The optimization formulas are different during the pretraining process for different pretraining models. In this paper, we pretrained deep neutral network by different pretraining models and hence investigated the difference between DBN and Stacked Denoising Autoencoder (SDA) when used as pretraining model. The experimental results show that DBN get a better initial
more » ... l. However the model converges to a relatively worse model after the finetuning process. Yet after pretrained by SDA for the second time the model converges to a better model if finetuned.
arXiv:1606.00540v1 fatcat:e4tnrji6kbagpk36yxy32bfykm

Pd(OAc)2/CuI-Catalyzed Tandem Reaction for Synthesis of Polysubstituted 3-Chalcogenylindoles

Ruiting Liu, Zhen Li, Shengke Wang, Xigeng Zhou
2019 Youji huaxue  
研究论文 Pd(OAc) 2 /CuI 分步催化合成多取代 3-烃硫(硒)基吲哚化合物 关键词 吲哚; Pd(OAc) 2 /CuI; 催化; 合成 Abstract Tandem Pd(OAc) 2 /CuI catalyzed coupling/cyclization/chalcogenylation reaction of gem-dibromovinylanilines with boronic  ...  by reaction of CuI with dichalcogenide, [17] on indole ring results in the formation of chalcogenonium cation (III).  ...  In addition, the reaction of HSC with air in the presence of CuI (0.10 equiv.) was carried out and 3-MeH 4 C 6 SSC 6 H 4 Me-3 (3c) was obtained in 93% (Scheme 2, d).  ... 
doi:10.6023/cjoc201904032 fatcat:7hju5q26ubhbff676yw3sfiyx4

A novel arabinose-inducible genetic operation system developed for Clostridium cellulolyticum

Jie Zhang, Ya-Jun Liu, Gu-Zhen Cui, Qiu Cui
2015 Biotechnology for Biofuels  
Cui et al. disrupted the MspI encoding gene in C. cellulolyticum to construct a cell chassis that requires no methylation of heterologous DNA [14] , and further developed a pyrFbased assistant system  ... 
doi:10.1186/s13068-015-0214-2 pmid:25763107 pmcid:PMC4355141 fatcat:rztauquskrgtjdsw5zlvza7bwq

Differentiating Benign from Malignant Bone Tumors Using Fluid-Fluid Level Features on Magnetic Resonance Imaging

Hong Yu, Jian-Ling Cui, Sheng-Jie Cui, Ying-Cai Sun, Feng-Zhen Cui
2014 Korean Journal of Radiology  
Objective: To analyze different fluid-fluid level features between benign and malignant bone tumors on magnetic resonance imaging (MRI). Materials and Methods: This study was approved by the hospital ethics committee. We retrospectively analyzed 47 patients diagnosed with benign (n = 29) or malignant (n = 18) bone tumors demonstrated by biopsy/surgical resection and who showed the intratumoral fluid-fluid level on pre-surgical MRI. The maximum length of the largest fluid-fluid level and the
more » ... o of the maximum length of the largest fluid-fluid level to the maximum length of a bone tumor in the sagittal plane were investigated for use in distinguishing benign from malignant tumors using the Mann-Whitney U-test and a receiver operating characteristic (ROC) analysis. Fluid-fluid level was categorized by quantity (multiple vs. single fluid-fluid level) and by T1-weighted image signal pattern (high/low, low/high, and undifferentiated), and the findings were compared between the benign and malignant groups using the χ 2 test. Results: The ratio of the maximum length of the largest fluid-fluid level to the maximum length of bone tumors in the sagittal plane that allowed statistically significant differentiation between benign and malignant bone tumors had an area under the ROC curve of 0.758 (95% confidence interval, 0.616-0.899). A cutoff value of 41.5% (higher value suggests a benign tumor) had sensitivity of 73% and specificity of 83%. Conclusion: The ratio of the maximum length of the largest fluid-fluid level to the maximum length of a bone tumor in the sagittal plane may be useful to differentiate benign from malignant bone tumors.
doi:10.3348/kjr.2014.15.6.757 pmid:25469087 pmcid:PMC4248631 fatcat:wvlzoanxbjgjnhsjum5w24j5vi

Graph Jigsaw Learning for Cartoon Face Recognition [article]

Yong Li, Lingjie Lao, Zhen Cui, Shiguang Shan, Jian Yang
2021 arXiv   pre-print
Zhen Cui received the Ph.D. degree from Institute of Computing Technology (ICT), Chinese Academy of Sciences in 2014.  ... 
arXiv:2107.06532v1 fatcat:lg5lntck2ncpnnrp3ldolx5xrm

Dual-Attention Graph Convolutional Network [article]

Xueya Zhang and Tong Zhang and Wenting Zhao and Zhen Cui and Jian Yang
2019 arXiv   pre-print
Graph convolutional networks (GCNs) have shown the powerful ability in text structure representation and effectively facilitate the task of text classification. However, challenges still exist in adapting GCN on learning discriminative features from texts due to the main issue of graph variants incurred by the textual complexity and diversity. In this paper, we propose a dual-attention GCN to model the structural information of various texts as well as tackle the graph-invariant problem through
more » ... embedding two types of attention mechanisms, i.e. the connection-attention and hop-attention, into the classic GCN. To encode various connection patterns between neighbour words, connection-attention adaptively imposes different weights specified to neighbourhoods of each word, which captures the short-term dependencies. On the other hand, the hop-attention applies scaled coefficients to different scopes during the graph diffusion process to make the model learn more about the distribution of context, which captures long-term semantics in an adaptive way. Extensive experiments are conducted on five widely used datasets to evaluate our dual-attention GCN, and the achieved state-of-the-art performance verifies the effectiveness of dual-attention mechanisms.
arXiv:1911.12486v1 fatcat:a5x4f3ypknhdrb7esmdgndpwzy

Shallow Attention Network for Polyp Segmentation [article]

Jun Wei, Yiwen Hu, Ruimao Zhang, Zhen Li, S.Kevin Zhou, Shuguang Cui
2021 arXiv   pre-print
Accurate polyp segmentation is of great importance for colorectal cancer diagnosis. However, even with a powerful deep neural network, there still exists three big challenges that impede the development of polyp segmentation. (i) Samples collected under different conditions show inconsistent colors, causing the feature distribution gap and overfitting issue; (ii) Due to repeated feature downsampling, small polyps are easily degraded; (iii) Foreground and background pixels are imbalanced,
more » ... to a biased training. To address the above issues, we propose the Shallow Attention Network (SANet) for polyp segmentation. Specifically, to eliminate the effects of color, we design the color exchange operation to decouple the image contents and colors, and force the model to focus more on the target shape and structure. Furthermore, to enhance the segmentation quality of small polyps, we propose the shallow attention module to filter out the background noise of shallow features. Thanks to the high resolution of shallow features, small polyps can be preserved correctly. In addition, to ease the severe pixel imbalance for small polyps, we propose a probability correction strategy (PCS) during the inference phase. Note that even though PCS is not involved in the training phase, it can still work well on a biased model and consistently improve the segmentation performance. Quantitative and qualitative experimental results on five challenging benchmarks confirm that our proposed SANet outperforms previous state-of-the-art methods by a large margin and achieves a speed about 72FPS.
arXiv:2108.00882v1 fatcat:k2aae5k5yvhwjdhp76lzwmicda
« Previous Showing results 1 — 15 out of 4,901 results