38,707 Hits in 1.8 sec

Region Mutual Information Loss for Semantic Segmentation [article]

Shuai Zhao, Yang Wang, Zheng Yang, Deng Cai
2019 arXiv   pre-print
Semantic segmentation is a fundamental problem in computer vision. It is considered as a pixel-wise classification problem in practice, and most segmentation models use a pixel-wise loss as their optimization riterion. However, the pixel-wise loss ignores the dependencies between pixels in an image. Several ways to exploit the relationship between pixels have been investigated, \eg, conditional random fields (CRF) and pixel affinity based methods. Nevertheless, these methods usually require
more » ... tional model branches, large extra memories, or more inference time. In this paper, we develop a region mutual information (RMI) loss to model the dependencies among pixels more simply and efficiently. In contrast to the pixel-wise loss which treats the pixels as independent samples, RMI uses one pixel and its neighbour pixels to represent this pixel. Then for each pixel in an image, we get a multi-dimensional point that encodes the relationship between pixels, and the image is cast into a multi-dimensional distribution of these high-dimensional points. The prediction and ground truth thus can achieve high order consistency through maximizing the mutual information (MI) between their multi-dimensional distributions. Moreover, as the actual value of the MI is hard to calculate, we derive a lower bound of the MI and maximize the lower bound to maximize the real value of the MI. RMI only requires a few extra computational resources in the training stage, and there is no overhead during testing. Experimental results demonstrate that RMI can achieve substantial and consistent improvements in performance on PASCAL VOC 2012 and CamVid datasets. The code is available at
arXiv:1910.12037v1 fatcat:b7k6yep3tjgt7cdqdea4s22s7u

Association of Radiotherapy for Rectal Cancer and Second Gynecological Malignant Neoplasms

Xu Guan, Ran Wei, Runkun Yang, Zhao Lu, Enrui Liu, Zhixun Zhao, Haipeng Chen, Ming Yang, Zheng Liu, Zheng Jiang, Xishan Wang
2021 JAMA Network Open  
Radiotherapy is a common treatment for rectal cancer, yet the risk of second gynecological malignant neoplasms (SGMNs) in patients with rectal cancer undergoing radiotherapy have not been adequately studied. To investigate the association between radiotherapy and the risk of individual types of SGMN in patients with rectal cancer and assess survival outcomes. A large population-based cohort study was designed to identify the risk of SGMNs in patients with rectal cancer diagnosed from January
more » ... 3 to December 2015. The statistical analysis was conducted from September 2019 to April 2020. The study was based on the 9 cancer registries of Surveillance, Epidemiology, and End Results database. A total of 20 142 female patients with rectal cancer in localized and regional stage were included. Receipt of neoadjuvant radiotherapy for rectal cancer. The development of an SGMN defined as any type of GMN occurring more than 5 years after the diagnosis of rectal cancer. The cumulative incidence of SGMNs was estimated by Fine-Gray competing risk regression. Poisson regression was used to evaluate the radiotherapy-associated risk for SGMNs in patients undergoing radiotherapy vs patients not undergoing radiotherapy. The Kaplan-Meier method was used to assess the survival outcomes of patients with SGMNs. Of 20 142 patients, 16 802 patients (83.4%) were White and the median age was 65 years (interquartile range, 54-74 years). A total of 5310 (34.3%) patients were treated with surgery and radiotherapy, and 14 832 (65.7%) patients were treated with surgery alone. The cumulative incidence of SGMNs during 30 years of follow-up was 4.53% among patients who received radiotherapy and 1.53% among patients who did not. In competing risk regression analysis, undergoing radiotherapy was associated with a higher risk of developing cancer of the uterine corpus (adjusted hazard ratio, 3.06; 95% CI, 2.14-4.37; P < .001) and ovarian cancer (adjusted hazard ratio, 2.08; 95% CI, 1.22-3.56; P = .007) compared with those who did not receive radiotherapy. The dynamic radiotherapy-associated risks (RR) for cancer of the uterine corpus significantly increased with increasing age at rectal cancer diagnosis (aged 20-49 years: adjusted RR, 0.79; 95% CI, 0.35-1.79; P = .57; aged 50-69 years: adjusted RR, 3.74; 95% CI, 2.63-5.32; P < .001; aged ≥70 years: adjusted RR, 5.13; 95% CI, 2.64-9.97; P < .001) and decreased with increasing latency since rectal cancer diagnosis (60-119 months: adjusted RR, 3.22; 95% CI, 2.12-4.87; P < .001; 120-239 months: adjusted RR, 2.72; 95% CI, 1.75-4.24; P < .001; 240-360 months: adjusted RR, 1.95; 95% CI, 0.67-5.66; P = .22), but the dynamic RR for ovarian cancer increased with increasing latency since rectal cancer diagnosis (60-119 months: adjusted RR, 0.70; 95% CI, 0.26-1.89; P = .48; 120-239 months: adjusted RR, 2.26; 95% CI, 1.09-4.70; P = .03; 240-360 months: adjusted RR, 11.84; 95% CI, 2.18-64.33; P = .004). The 10-year overall survival among patients with radiotherapy-associated cancer of the uterine corpus was significantly lower than that among matched patients with primary cancer of the uterine corpus (21.5% vs 33.6%; P = .01). Radiotherapy for rectal cancer was associated with an increased risk of cancer of the uterine corpus and ovarian cancer. Special attention should be paid to reduce radiotherapy-associated SGMNs and improve their prognosis.
doi:10.1001/jamanetworkopen.2020.31661 pmid:33416884 pmcid:PMC7794669 fatcat:2i7jlcadtzeujimxrdzpkymnbu

Maxwell's demon and Smoluchowski's trap door

Jianzhou Zheng, Xiao Zheng, Yang Zhao, Yang Xie, ChiYung Yam, GuanHua Chen, Qing Jiang, Allen T. Chwang
2007 Physical Review E  
As time goes on, the particles in the left compartment will have a higher temperature ͑and a higher particle density͒ than that of the right, and the *Electronic address: † Electronic  ... 
doi:10.1103/physreve.75.041109 pmid:17500867 fatcat:4qrujla5gvazjnac7cw2xtgndy

Barter Exchange via Friends' Friends [article]

Yue Zheng, Tianyi Yang, Wen Zhang, Dengji Zhao
2020 arXiv   pre-print
This idea is inspired by the work of Li et al. 2017 and Zhao et al. 2018 .  ... 
arXiv:2010.04933v1 fatcat:xvze5hugqvajjhg5hcqlf625ei

Multi-modal Graph Learning for Disease Prediction [article]

Shuai Zheng, Zhenfeng Zhu, Zhizhe Liu, Zhenyu Guo, Yang Liu, Yuchen Yang, Yao Zhao
2022 arXiv   pre-print
Benefiting from the powerful expressive capability of graphs, graph-based approaches have been popularly applied to handle multi-modal medical data and achieved impressive performance in various biomedical applications. For disease prediction tasks, most existing graph-based methods tend to define the graph manually based on specified modality (e.g., demographic information), and then integrated other modalities to obtain the patient representation by Graph Representation Learning (GRL).
more » ... , constructing an appropriate graph in advance is not a simple matter for these methods. Meanwhile, the complex correlation between modalities is ignored. These factors inevitably yield the inadequacy of providing sufficient information about the patient's condition for a reliable diagnosis. To this end, we propose an end-to-end Multi-modal Graph Learning framework (MMGL) for disease prediction with multi-modality. To effectively exploit the rich information across multi-modality associated with the disease, modality-aware representation learning is proposed to aggregate the features of each modality by leveraging the correlation and complementarity between the modalities. Furthermore, instead of defining the graph manually, the latent graph structure is captured through an effective way of adaptive graph learning. It could be jointly optimized with the prediction model, thus revealing the intrinsic connections among samples. Our model is also applicable to the scenario of inductive learning for those unseen data. An extensive group of experiments on two disease prediction tasks demonstrates that the proposed MMGL achieves more favorable performance. The code of MMGL is available at .
arXiv:2203.05880v1 fatcat:cr47jpdbk5edha2qdkbtxb3xjq

FAMLP: A Frequency-Aware MLP-Like Architecture For Domain Generalization [article]

Kecheng Zheng, Yang Cao, Kai Zhu, Ruijing Zhao, Zheng-Jun Zha
2022 arXiv   pre-print
MLP-like models built entirely upon multi-layer perceptrons have recently been revisited, exhibiting the comparable performance with transformers. It is one of most promising architectures due to the excellent trade-off between network capability and efficiency in the large-scale recognition tasks. However, its generalization performance to heterogeneous tasks is inferior to other architectures (e.g., CNNs and transformers) due to the extensive retention of domain information. To address this
more » ... oblem, we propose a novel frequency-aware MLP architecture, in which the domain-specific features are filtered out in the transformed frequency domain, augmenting the invariant descriptor for label prediction. Specifically, we design an adaptive Fourier filter layer, in which a learnable frequency filter is utilized to adjust the amplitude distribution by optimizing both the real and imaginary parts. A low-rank enhancement module is further proposed to rectify the filtered features by adding the low-frequency components from SVD decomposition. Finally, a momentum update strategy is utilized to stabilize the optimization to fluctuation of model parameters and inputs by the output distillation with weighted historical states. To our best knowledge, we are the first to propose a MLP-like backbone for domain generalization. Extensive experiments on three benchmarks demonstrate significant generalization performance, outperforming the state-of-the-art methods by a margin of 3%, 4% and 9%, respectively.
arXiv:2203.12893v1 fatcat:uqg3eup6rjhk5om5fam7zvg4vy

Transcranial brain atlas

Xiang Xiao, Xiaoting Yu, Zong Zhang, Yang Zhao, Yihan Jiang, Zheng Li, Yihong Yang, Chaozhe Zhu
2018 Science Advances  
We introduce here the concept of a transcranial brain atlas (TBA), a new kind of brain atlas specialized for transcranial techniques. A TBA is a probabilistic mapping from scalp space to atlas label space, relating scalp locations to anatomical, functional, network, genetic, or other labels. TBAs offer a new way to integrate and present structural and functional organization of the brain and allow previously subsurface and invisible atlas labels visible on the scalp surface to accurately guide
more » ... he placement of transcranial devices directly on the scalp surface in a straightforward, visual manner. We present here a framework for building TBAs that includes (i) a new, continuous proportional coordinate system devised for the scalp surface to allow standardized specification of scalp positions; (ii) a high-resolution, large sample-based (114-participant) mapping from scalp space to brain space to accurately and reliably describe human cranio-cortical correspondence; and (iii) a two-step Markov chain to combine the probabilistic scalp-brain mapping with a traditional brain atlas, bringing atlas labels to the scalp surface. We assessed the reproducibility (consistency of TBAs generated from different groups) and predictiveness (prediction accuracy of labels for individuals without brain images) of the TBAs built via our framework. Moreover, we present an application of TBAs to a functional near-infrared spectroscopy finger-tapping experiment, illustrating the utility and benefits of TBAs in transcranial studies. Our results demonstrate that TBAs can support ongoing efforts to map the human brain using transcranial techniques, just as traditional brain atlases have supported magnetic resonance imaging and positron emission tomography studies.
doi:10.1126/sciadv.aar6904 pmid:30191174 pmcid:PMC6124906 fatcat:shnggwa4lfcmrntwukjtv4uteu

An Iterative Algorithm For Klda Classifier

D.N. Zheng, J.X. Wang, Y.N. Zhao, Z.H. Yang
2007 Zenodo  
The Linear discriminant analysis (LDA) can be generalized into a nonlinear form - kernel LDA (KLDA) expediently by using the kernel functions. But KLDA is often referred to a general eigenvalue problem in singular case. To avoid this complication, this paper proposes an iterative algorithm for the two-class KLDA. The proposed KLDA is used as a nonlinear discriminant classifier, and the experiments show that it has a comparable performance with SVM.
doi:10.5281/zenodo.1059321 fatcat:xlzc2wksrjd2ldkjiqmj6ebbzy

Reliable and Efficient Autonomous Driving: the Need for Heterogeneous Vehicular Networks [article]

Kan Zheng, Qiang Zheng, Haojun Yang, Long Zhao, Lu Hou, Periklis Chatzimisios
2015 arXiv   pre-print
Long Zhao received the Ph.D. degree from Beijing University of Posts and Telecommunications, Beijing, China, in 2015, where he is currently a lecturer.  ... 
arXiv:1510.06607v1 fatcat:6vrdjsw7mrgiro6yr6dx7wnt3q


Lei Yang, Zengbin Zhang, Wei Hou, Ben Y. Zhao, Haitao Zheng
2011 Computer communication review  
Proliferation and innovation of wireless technologies require significant amounts of radio spectrum. Recent policy reforms by the FCC are paving the way by freeing up spectrum for a new generation of frequency-agile wireless devices based on software defined radios (SDRs). But despite recent advances in SDR hardware, research on SDR MAC protocols or applications requires an experimental platform for managing physical access. We introduce Papyrus, a software platform for wireless researchers to
more » ... evelop and experiment dynamic spectrum systems using currently available SDR hardware. Papyrus provides two fundamental building blocks at the physical layer: flexible non-contiguous frequency access and simple and robust frequency detection. Papyrus allows researchers to deploy and experiment new MAC protocols and applications on USRP GNU Radio, and can also be ported to other SDR platforms. We demonstrate the use of Papyrus using Jello, a distributed MAC overlay for high-bandwidth media streaming applications and Ganache, a SDR layer for adaptable guardband configuration. Full implementations of Papyrus and Jello are publicly available.
doi:10.1145/1925861.1925866 fatcat:avaz4l2b3bcyxoafddysiivbii

Saliency driven vasculature segmentation with infinite perimeter active contour model

Yitian Zhao, Jingliang Zhao, Jian Yang, Yonghuai Liu, Yifan Zhao, Yalin Zheng, Likun Xia, Yongtian Wang
2017 Neurocomputing  
Jian YANG received his Ph.D. degree in optical engineering from Beijing Institute of Technology in 2007.  ...  Most recently, Zhao et al. [49, 50] adapted the saliency concept to detecting abnormalities characteristic of malarial retinopathy (MR).  ... 
doi:10.1016/j.neucom.2016.07.077 fatcat:dzd4g34b6jd35hqugwoviqxcxa

Laboratory diagnosis and monitoring the viral shedding of 2019-nCoV infections [article]

Yang Yang, Minghui Yang, Chenguang Shen, Fuxiang Wang, Jing Yuan, Jinxiu Li, Mingxia Zhang, Zhaoqin Wang, Li Xing, Jinli Wei, Ling Peng, Gary Wong (+10 others)
2020 medRxiv   pre-print
The outbreak of novel coronavirus pneumonia (NCP) caused by 2019-nCoV spread rapidly, and elucidation the diagnostic accuracy of different respiratory specimens is crucial for the control and treatment of this diseases. Methods: Respiratory samples including nasal swabs, throat swabs, sputum and bronchoalveolar lavage fluid (BALF) were collected from Guangdong CDC confirmed NCP patients, and viral RNAs were detected using a CFDA approved detection kit. Results were analyzed in combination with
more » ... ample collection date and clinical information. Finding: Except for BALF, the sputum possessed the highest positive rate (74.4%~88.9%), followed by nasal swabs (53.6%~73.3%) for both severe and mild cases during the first 14 days after illness onset (d.a.o). For samples collected ≥ 15 d.a.o, sputum and nasal swabs still possessed a high positive rate ranging from 42.9%~61.1%. The positive rate of throat swabs collected ≥ 8 d.a.o was low, especially in samples from mild cases. Viral RNAs could be detected in all the lower respiratory tract of severe cases, but not the mild cases. CT scan of cases 02, 07 and 13 showed typical viral pneumonia with ground glass opacity, while no viral RNAs were detected in first three or all the upper respiratory samples. Interpretation: Sputum is most accurate for laboratory diagnosis of NCP, followed by nasal swabs. Detection of viral RNAs in BLAF is necessary for diagnosis and monitoring of viruses in severe cases. CT scan could serve as an important make up for the diagnosis of NCP. Funding National Science and Technology Major Project, Sanming Project of Medicine and China Postdoctoral Science Foundation.
doi:10.1101/2020.02.11.20021493 fatcat:m6so6oilynhqxgmge5buletzki

Reduction of aliasing artifact in Micro CT

Luo Zhao-Yang, Yang Xiao-Quan, Meng Yuan-Zheng, Deng Yong
2010 Wuli xuebao  
doi:10.7498/aps.59.8237 fatcat:54fy6sx7uvamxd7n3bxn2fgeii

Technical report on Conversational Question Answering [article]

Ying Ju, Fubang Zhao, Shijie Chen, Bowen Zheng, Xuefeng Yang, Yunfeng Liu
2019 arXiv   pre-print
., 2018) , XLNET (Yang et al., 2019) and RoBERTa (Liu et al., 2019) , have brought significant performance gains to many NLP tasks, including machine reading comprehension.  ... 
arXiv:1909.10772v1 fatcat:5zr6jnvo4rhvvcjrz3nhfzd3wm

Tortoise coordinate and Hawking effect in the Kinnersley spacetime [article]

Jian Yang, Zheng Zhao, Wenbiao Liu
2010 arXiv   pre-print
Acknowledgement One of the authors, Jian Yang, would like to thank Dr. Shiwei Zhou and Dr. Xianming Liu for their helpful discussions .  ...  In 1990's, Z Zhao, X. X. Dai and Z. Q. Luo improved Damour-Ruffini method to study Hawking effect from some dynamical black holes.  ... 
arXiv:1003.2686v1 fatcat:fsj7srrvf5eazjc3rhehyw4hem
« Previous Showing results 1 — 15 out of 38,707 results