The Chinese University of Hong Kong

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 & Computer science. The organization has 43411 authors who have published 93672 publications receiving 3066651 citations.

Accelerating the Super-Resolution Convolutional Neural Network

Abstract: As a successful deep model applied in image super-resolution (SR), the Super-Resolution Convolutional Neural Network (SRCNN) has demonstrated superior performance to the previous hand-crafted models either in speed and restoration quality. However, the high computational cost still hinders it from practical usage that demands real-time performance (24 fps). In this paper, we aim at accelerating the current SRCNN, and propose a compact hourglass-shape CNN structure for faster and better SR. We re-design the SRCNN structure mainly in three aspects. First, we introduce a deconvolution layer at the end of the network, then the mapping is learned directly from the original low-resolution image (without interpolation) to the high-resolution one. Second, we reformulate the mapping layer by shrinking the input feature dimension before mapping and expanding back afterwards. Third, we adopt smaller filter sizes but more mapping layers. The proposed model achieves a speed up of more than 40 times with even superior restoration quality. Further, we present the parameter settings that can achieve real-time performance on a generic CPU while still maintaining good performance. A corresponding transfer strategy is also proposed for fast training and testing across different upscaling factors.

Rolling Guidance Filter

Abstract: Images contain many levels of important structures and edges. Compared to masses of research to make filters edge preserving, finding scale-aware local operations was seldom addressed in a practical way, albeit similarly vital in image processing and computer vision. We propose a new framework to filter images with the complete control of detail smoothing under a scale measure. It is based on a rolling guidance implemented in an iterative manner that converges quickly. Our method is simple in implementation, easy to understand, fully extensible to accommodate various data operations, and fast to produce results. Our implementation achieves realtime performance and produces artifact-free results in separating different scale structures. This filter also introduces several inspiring properties different from previous edge-preserving ones.

Risk estimation for hepatocellular carcinoma in chronic hepatitis B (REACH-B): development and validation of a predictive score

Abstract: Summary Background Therapy for chronic hepatitis B reduces the risk of progressing to hepatocellular carcinoma (HCC); however, there is no suitable and accurate means to assess risk. This study aimed to develop and validate a simple scoring system to predict HCC risk in patients with chronic hepatitis B. Methods The development cohort consisted of 3584 patients without cirrhosis from the community-based Taiwanese REVEAL-HBV study (of whom 131 developed HCC during follow-up), and a validation cohort of 1505 patients from three hospitals in Hong Kong and South Korea (of whom 111 developed HCC during follow-up). We used Cox multivariate proportional hazards model to predict risk of HCC at 3, 5, and 10 years. Variables included in the risk score were sex, age, serum alanine aminotransferase concentration, HBeAg status, and serum HBV DNA level. We calculated the area under receiver operating curve (AUROC) and calibration of predicted and observed HCC risk. Findings A 17-point risk score was developed, with HCC risk ranging from 0·0% to 23·6% at 3 years, 0·0% to 47·4% at 5 years, and 0·0% to 81·6% at 10 years for patients with the lowest and highest HCC risk, respectively. AUROCs to predict risk were 0·811 (95% CI 0·790–0·831) at 3 years, 0·796 (0·775–0·816) at 5 years, and 0·769 (0·747–0·790) at 10 years in the validation cohort, and 0·902 (0·884–0·918), 0·783 (0·759–0·806), and 0·806 (0·783–0·828), respectively, after exclusion of 277 patients in the validation cohort with cirrhosis. Predicted risk was well calibrated with Kaplan-Meier observed HCC risk. Interpretation A simple-to-use risk score that uses baseline clinical variables was developed and validated. The score accurately estimates the risk of developing HCC at 3, 5, and 10 years in patients with chronic hepatitis B. Clinicians can use this score to assess risk of HCC in patients with chronic hepatitis B and subsequently make evidence-based decisions about their clinical management. Funding The Academia Sinica; the National Health Research Institute, Taiwan; and Bristol-Myers Squibb.

Restrictive versus Liberal Fluid Therapy for Major Abdominal Surgery

Abstract: Background Guidelines to promote the early recovery of patients undergoing major surgery recommend a restrictive intravenous-fluid strategy for abdominal surgery. However, the supporting evidence is limited, and there is concern about impaired organ perfusion. Methods In a pragmatic, international trial, we randomly assigned 3000 patients who had an increased risk of complications while undergoing major abdominal surgery to receive a restrictive or liberal intravenous-fluid regimen during and up to 24 hours after surgery. The primary outcome was disability-free survival at 1 year. Key secondary outcomes were acute kidney injury at 30 days, renal-replacement therapy at 90 days, and a composite of septic complications, surgical-site infection, or death. Results During and up to 24 hours after surgery, 1490 patients in the restrictive fluid group had a median intravenous-fluid intake of 3.7 liters (interquartile range, 2.9 to 4.9), as compared with 6.1 liters (interquartile range, 5.0 to 7.4) in 149...