Guan Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Guan Wu has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
G. Zhang
2 shared publications- Ruijing Xiao
Department of Pathophysiology, School of Basic Medical Sciences, Wuhan University, Wuhan 430071, China
2 shared publications X. Wang
2 shared publicationsD. Long
1 shared publicationDragan Golijanin
Department of Urology, University of Rochester Medical Center, Rochester, NY 14642, USA
1 shared publicationDragan J. Golijanin
Department of Urology, University of Rochester Medical Center, Rochester, USA
1 shared publication
Latest Publications
- Open Access
- 26 July 2023
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Graphics
- Vol. 42 (4) , 1-16
ScanBot: Autonomous Reconstruction via Deep Reinforcement Learning
Autoscanning of an unknown environment is the key to many AR/VR and robotic applications. However, autonomous reconstruction with both high efficiency and quality remains a challenging problem. In this work, we propose a reconstruction-oriented autoscanning approach, called ScanBot, which utilizes hierarchical deep reinforcement learning techniques for global region-of-interest (ROI) planning to improve the scanning efficiency and local next-best-view (NBV) planning to enhance the reconstruction quality. Given the partially reconstructed scene, the global policy designates an ROI with insufficient exploration or reconstruction. The local policy is then applied to refine the reconstruction quality of objects in this region by planning and scanning a series of NBVs. A novel mixed 2D-3D representation is designed for these policies, where a 2D quality map with tailored quality channels encoding the scanning progress is consumed by the global policy, and a coarse-to-fine 3D volumetric representation that embodies both local environment and object completeness is fed to the local policy. These two policies iterate until the whole scene has been completely explored and scanned. To speed up the learning of complex environmental dynamics and enhance the agent's memory for spatial-temporal inference, we further introduce two novel auxiliary learning tasks to guide the training of our global policy. Thorough evaluations and comparisons are carried out to show the feasibility of our proposed approach and its advantages over previous methods. Code and data are available at https://github.com/HezhiCao/Scanbot.
- 26 July 2023
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Graphics
- Vol. 42 (4) , 1-16
Open AccessScanBot: Autonomous Reconstruction via Deep Reinforcement Learning
Autoscanning of an unknown environment is the key to many AR/VR and robotic applications. However, autonomous reconstruction with both high efficiency and quality remains a challenging problem. In this work, we propose a reconstruction-oriented autoscanning approach, called ScanBot, which utilizes hierarchical deep reinforcement learning techniques for global region-of-interest (ROI) planning to improve the scanning efficiency and local next-best-view (NBV) planning to enhance the reconstruction quality. Given the partially reconstructed scene, the global policy designates an ROI with insufficient exploration or reconstruction. The local policy is then applied to refine the reconstruction quality of objects in this region by planning and scanning a series of NBVs. A novel mixed 2D-3D representation is designed for these policies, where a 2D quality map with tailored quality channels encoding the scanning progress is consumed by the global policy, and a coarse-to-fine 3D volumetric representation that embodies both local environment and object completeness is fed to the local policy. These two policies iterate until the whole scene has been completely explored and scanned. To speed up the learning of complex environmental dynamics and enhance the agent's memory for spatial-temporal inference, we further introduce two novel auxiliary learning tasks to guide the training of our global policy. Thorough evaluations and comparisons are carried out to show the feasibility of our proposed approach and its advantages over previous methods. Code and data are available at https://github.com/HezhiCao/Scanbot.
- 15 March 2019
- Published by Elsevier in European Urology Supplements
- Vol. 18 (1) , e2153
Innovation and development of laparoendoscopic nephroureterectomy: One-center experience
- 15 March 2019
- Published by Elsevier in European Urology Supplements
- Vol. 18 (1) , e2153
Innovation and development of laparoendoscopic nephroureterectomy: One-center experience
- 1 July 2013
- Published by Elsevier in Urologic Oncology: Seminars and Original Investigations
- Vol. 31 (5) , 656-663
Non-bladder cancer mortality in patients with urothelial cancer of the bladder
- 1 July 2013
- Published by Elsevier in Urologic Oncology: Seminars and Original Investigations
- Vol. 31 (5) , 656-663
Non-bladder cancer mortality in patients with urothelial cancer of the bladder
- 1 March 2013
- Published by Elsevier in European Urology Supplements
- Vol. 12 (1) , e948
948 Hybrid transvaginal natural orifice transluminal endoscopic surgery (NOTES) nephrectomy: Our experience
- 1 March 2013
- Published by Elsevier in European Urology Supplements
- Vol. 12 (1) , e948
948 Hybrid transvaginal natural orifice transluminal endoscopic surgery (NOTES) nephrectomy: Our experience
- 10 June 2011
- Published by Springer Nature in Journal of Robotic Surgery
- Vol. 6 (2) , 139-147
Robot-assisted partial nephrectomy: analysis of the first 100 cases from a single institution
Robot-assisted partial nephrectomy (RAPN) is an alternative to open and laparoscopic partial nephrectomy for small renal tumors. Our objectives were to report our experience and short-term outcomes from the first 100 cases of robot-assisted partial nephrectomy (RAPN) performed at a single institution, as well as to evaluate the effect of the learning curve and identify any factors associated with adverse perioperative outcomes. Patient records of the first 100 RAPN cases performed by three surgeons between October 2007 and March 2010 were retrospectively reviewed. The cases were divided into two groups to analyze a possible learning curve effect. Group 1 consisted of the first half (chronologically) of the cases performed by each surgeon, and Group 2 consisted of the second half. For the entire series, the median warm ischemia time was 24 min (range 11–49), mean length of follow-up was 13.4 months, and the median postoperative change in glomerular filtration rate (GFR) was −6.6 mL/min/1.73 m2. Three patients had microscopically positive margins on final pathology, three intraoperative complications occurred, and 13 postoperative complications were recorded (10 Clavien grade IIIa or less). Median operative time was significantly longer in Group 1 (193 min) than in Group 2 (165 min, P = 0.003). Multivariate analysis identified male gender and cases done in Group 1 to be associated with increased operative time, while male gender and higher nephrometry scores were associated with increased blood loss. Tumor characteristics associated with greater reductions in GFR included higher nephrometry scores, endophytic tumors, and hilar tumors. In conclusion, RAPN appears to be safe and the major effect of the learning curve appears to be on operative time. Warm ischemia times are sufficiently low to prevent significant renal impairment, while male gender and higher nephrometry scores may be predictors of longer operative times and more intraoperative blood loss. Overall operative time decreased with increasing case volume, although this was not uniform among the three surgeons in the study. Further longitudinal study is necessary to establish oncologic outcomes.
- 10 June 2011
- Published by Springer Nature in Journal of Robotic Surgery
- Vol. 6 (2) , 139-147
Robot-assisted partial nephrectomy: analysis of the first 100 cases from a single institution
Robot-assisted partial nephrectomy (RAPN) is an alternative to open and laparoscopic partial nephrectomy for small renal tumors. Our objectives were to report our experience and short-term outcomes from the first 100 cases of robot-assisted partial nephrectomy (RAPN) performed at a single institution, as well as to evaluate the effect of the learning curve and identify any factors associated with adverse perioperative outcomes. Patient records of the first 100 RAPN cases performed by three surgeons between October 2007 and March 2010 were retrospectively reviewed. The cases were divided into two groups to analyze a possible learning curve effect. Group 1 consisted of the first half (chronologically) of the cases performed by each surgeon, and Group 2 consisted of the second half. For the entire series, the median warm ischemia time was 24 min (range 11–49), mean length of follow-up was 13.4 months, and the median postoperative change in glomerular filtration rate (GFR) was −6.6 mL/min/1.73 m2. Three patients had microscopically positive margins on final pathology, three intraoperative complications occurred, and 13 postoperative complications were recorded (10 Clavien grade IIIa or less). Median operative time was significantly longer in Group 1 (193 min) than in Group 2 (165 min, P = 0.003). Multivariate analysis identified male gender and cases done in Group 1 to be associated with increased operative time, while male gender and higher nephrometry scores were associated with increased blood loss. Tumor characteristics associated with greater reductions in GFR included higher nephrometry scores, endophytic tumors, and hilar tumors. In conclusion, RAPN appears to be safe and the major effect of the learning curve appears to be on operative time. Warm ischemia times are sufficiently low to prevent significant renal impairment, while male gender and higher nephrometry scores may be predictors of longer operative times and more intraoperative blood loss. Overall operative time decreased with increasing case volume, although this was not uniform among the three surgeons in the study. Further longitudinal study is necessary to establish oncologic outcomes.