Institution
The Chinese University of Hong Kong
Education•Hong Kong, China•
About: The Chinese University of Hong Kong is a education organization based out in Hong Kong, China. It is known for research contribution in the topics: Population & Cancer. The organization has 43411 authors who have published 93672 publications receiving 3066651 citations.
Topics: Population, Cancer, Poison control, Randomized controlled trial, China
Papers published on a yearly basis
Papers
More filters
•
TL;DR: This paper reviews deep SOD algorithms from different perspectives, including network architecture, level of supervision, learning paradigm, and object-/instance-level detection, and looks into the generalization and difficulty of existing SOD datasets.
Abstract: As an essential problem in computer vision, salient object detection (SOD) has attracted an increasing amount of research attention over the years. Recent advances in SOD are predominantly led by deep learning-based solutions (named deep SOD). To enable in-depth understanding of deep SOD, in this paper, we provide a comprehensive survey covering various aspects, ranging from algorithm taxonomy to unsolved issues. In particular, we first review deep SOD algorithms from different perspectives, including network architecture, level of supervision, learning paradigm, and object-/instance-level detection. Following that, we summarize and analyze existing SOD datasets and evaluation metrics. Then, we benchmark a large group of representative SOD models, and provide detailed analyses of the comparison results. Moreover, we study the performance of SOD algorithms under different attribute settings, which has not been thoroughly explored previously, by constructing a novel SOD dataset with rich attribute annotations covering various salient object types, challenging factors, and scene categories. We further analyze, for the first time in the field, the robustness of SOD models to random input perturbations and adversarial attacks. We also look into the generalization and difficulty of existing SOD datasets. Finally, we discuss several open issues of SOD and outline future research directions.
428 citations
••
TL;DR: This systematic review does not support the traditional idea that ephedrine is the preferred choice for the management of maternal hypotension during spinal anesthesia for elective cesarean delivery in healthy, nonlaboring women.
Abstract: This quantitative systematic review compared the efficacy and safety of ephedrine with phenylephrine for the prevention and treatment of hypotension during spinal anesthesia for cesarean delivery. Seven randomized controlled trials (n 292) were identified after a systematic search of electronic databases (MEDLINE, EMBASE, The Cochrane Controlled Trials Registry), published articles, and contact with authors. Outcomes assessed were maternal hypotension, hypertension and bradycardia, and neonatal umbilical cord blood pH values and Apgar scores. For the management (prevention and treatment) of maternal hypotension, there was no difference between phenylephrine and ephedrine (relative risk [RR] of 1.00; 95% confidence interval [CI], 0.96 –1.06). Maternal bradycardia was more likely to occur with phenylephrine than with ephedrine (RR of 4.79; 95% CI, 1.47–15.60). Women given phenylephrine had neonates with higher umbilical arterial pH values than those given ephedrine (weighted mean difference of 0.03; 95% CI, 0.02– 0.04). There was no difference between the two vasopressors in the incidence of true fetal acidosis (umbilical arterial pH value of 7.2; RR of 0.78; 95% CI, 0.16 –3.92) or Apgar score of7 at 1 and 5 min. This systematic review does not support the traditional idea that ephedrine is the preferred choice for the management of maternal hypotension during spinal anesthesia for elective cesarean delivery in healthy, nonlaboring women. (Anesth Analg 2002;94:920 –6)
427 citations
••
18 May 2018TL;DR: This paper presents an alternative network that attains performance on par with FlowNet2 on the challenging Sintel final pass and KITTI benchmarks, while being 30 times smaller in the model size and 1.36 times faster in the running speed.
Abstract: FlowNet2 [14], the state-of-the-art convolutional neural network (CNN) for optical flow estimation, requires over 160M parameters to achieve accurate flow estimation. In this paper we present an alternative network that attains performance on par with FlowNet2 on the challenging Sintel final pass and KITTI benchmarks, while being 30 times smaller in the model size and 1.36 times faster in the running speed. This is made possible by drilling down to architectural details that might have been missed in the current frameworks: (1) We present a more effective flow inference approach at each pyramid level through a lightweight cascaded network. It not only improves flow estimation accuracy through early correction, but also permits seamless incorporation of descriptor matching in our network. (2) We present a novel flow regularization layer to ameliorate the issue of outliers and vague flow boundaries by using a feature-driven local convolution. (3) Our network owns an effective structure for pyramidal feature extraction and embraces feature warping rather than image warping as practiced in FlowNet2. Our code and trained models are available at github.com/twhui/LiteFlowNet.
427 citations
••
TL;DR: NAFLD is found in over a quarter of the general adult Chinese population, but the proportion of patients with advanced fibrosis is low and Modest alcohol consumption does not increase the risk of fatty liver or liver fibrosis.
Abstract: Objective Knowledge of the epidemiology of non-alcoholic fatty liver disease (NAFLD) is incomplete because liver biopsy cannot be performed on the general population to assess disease severity. New non-invasive tests allow accurate and safe assessment in healthy individuals. The aim of this study was to examine the prevalence of NAFLD and advanced fibrosis in the general Hong Kong Chinese population. Methods Subjects were recruited from the community by random selection from the government census database. Liver fat and fibrosis were assessed by proton-magnetic resonance spectroscopy and transient elastography, respectively. Results Overall, 264 of 922 (28.6%) subjects had intrahepatic triglyceride content ≥5%. Excluding 12 subjects with significant alcohol consumption, the population prevalence of NAFLD was 27.3% (95% CI 24.5% to 30.2%). Each component of the metabolic syndrome increased the risk of fatty liver in a dose-dependent manner (prevalence of 4.5% in subjects without any component and 80.0% in those with all five components). 8 (3.7%) patients with fatty liver had liver stiffness ≥9.6 kPa, a level suggestive of advanced fibrosis. Body mass index and alanine aminotransferase level were independent factors associated with liver stiffness. Together with other clinical prediction scores, the estimated prevalence of advanced fibrosis in patients with fatty liver in the community was Conclusion NAFLD is found in over a quarter of the general adult Chinese population, but the proportion of patients with advanced fibrosis is low. Modest alcohol consumption does not increase the risk of fatty liver or liver fibrosis.
426 citations
•
TL;DR: InterFaceGAN as discussed by the authors explores the disentanglement between various semantics and manage to decouple some entangled semantics with subspace projection, leading to more precise control of facial attributes, including gender, age, expression, and the presence of eyeglasses.
Abstract: Despite the recent advance of Generative Adversarial Networks (GANs) in high-fidelity image synthesis, there lacks enough understanding of how GANs are able to map a latent code sampled from a random distribution to a photo-realistic image. Previous work assumes the latent space learned by GANs follows a distributed representation but observes the vector arithmetic phenomenon. In this work, we propose a novel framework, called InterFaceGAN, for semantic face editing by interpreting the latent semantics learned by GANs. In this framework, we conduct a detailed study on how different semantics are encoded in the latent space of GANs for face synthesis. We find that the latent code of well-trained generative models actually learns a disentangled representation after linear transformations. We explore the disentanglement between various semantics and manage to decouple some entangled semantics with subspace projection, leading to more precise control of facial attributes. Besides manipulating gender, age, expression, and the presence of eyeglasses, we can even vary the face pose as well as fix the artifacts accidentally generated by GAN models. The proposed method is further applied to achieve real image manipulation when combined with GAN inversion methods or some encoder-involved models. Extensive results suggest that learning to synthesize faces spontaneously brings a disentangled and controllable facial attribute representation.
426 citations
Authors
Showing all 43993 results
Name | H-index | Papers | Citations |
---|---|---|---|
Michael Marmot | 193 | 1147 | 170338 |
Jing Wang | 184 | 4046 | 202769 |
Jiaguo Yu | 178 | 730 | 113300 |
Yang Yang | 171 | 2644 | 153049 |
Mark Gerstein | 168 | 751 | 149578 |
Gang Chen | 167 | 3372 | 149819 |
Jun Wang | 166 | 1093 | 141621 |
Jean Louis Vincent | 161 | 1667 | 163721 |
Wei Zheng | 151 | 1929 | 120209 |
Rui Zhang | 151 | 2625 | 107917 |
Ben Zhong Tang | 149 | 2007 | 116294 |
Kypros H. Nicolaides | 147 | 1302 | 87091 |
Thomas S. Huang | 146 | 1299 | 101564 |
Galen D. Stucky | 144 | 958 | 101796 |
Joseph J.Y. Sung | 142 | 1240 | 92035 |