Jing-yu Cui  

Education

2008.9 - Present

• Department of Electrical Engineering, Stanford University
• GPA: 4.20

2005.9 - 2008.7

• Master Candidate, Tsinghua University
• Advised by Professor Changshui Zhang
• Major in Computer Vision, Human Computer Interaction, Machine Learning, Information Retrieval.
• GPA: Overall: 91.6/100, Major: 92.1/100, Rank: 1/125

2001.9 - 2005.7

• B.E., Tsinghua University
• Major in Automatic Control Theory, Image Processing, Computer Vision.
• Admitted by Tsinghua with exemption from the College Entrance Examination for excellence in national Olympiad on mathematics, physics.
• GPA: Overall: 91.1/100, Major: 92.0/100, Rank: 4/225
  

1998.9 - 2001.7

• South Ocean International School, Shan xi, China.
  

Research

Interests

I am interested in practical computer vision and machine learning algorithms for robust and easy-to-use real life applications. My research targets at novel algorithms following human intuition, proper ways of incorporating human prior knowledge, and easy interactive scenarios of leveraging human cognitive power. I have experience in a broad range of topics in the field of computer vision, human-computer interaction, machine learning, computer graphics, and medical imaging, with an emphasize on: practical computer vision, interactive image organization, large scale image retrieval, image segmentation, machine learning algorithms in image retrieval, image/video compression, and massively parallel real-time dynamic medical imaging.

Current Work

I am currently working on machine learning and optimization methods for medical image analysis, especially for Positron Emission Tomography (PET), with an emphasize on massively parallel, real-time, dynamic imaging.

Past Work

	Massively Parallel Positron Emission Tomography (PET) Stanford, 2009.10-2010.8 Designed a parallel algorithm for computing line projection operations for fully 3-D PET image reconstruction. Implemented the list-mode OSEM algorithm on Graphics Processing Unit (GPU) using the Compute Unified Device Architecture (CUDA) framework. Increased the speed of calculation by 188 folds compared with the CPU version without changing the image quality. Published a book chapter (2nd author) and presented 2 conference posters.
	Automated CT Image Retrieval based on Liver Lesion Similarity Stanford, 2009.2-2009.10 Designed an algorithm based on AdaBoost and LASSO to learn an optimal way of combining heterogeneous features for similarity ranking of liver lesion CT images. Implemented a system to automatically test the performance of medical image retrieval systems. Coauthored a journal article and a conference abstract.
	Image/Video Compression based on Texture Analysis Google, 2009.6-2009.9 Designed and implemented a novel image/video compression algorithm using texture analysis. Improved over state-of-the-art method JPEG by 10% in terms of storage and bandwidth, and 0.5dB in terms of image/video quality. Tested on a large set of images/videos in MapReduce framework to demonstrate the consistent improvement of the algorithm. Filed an US Patent as the first inventor. Published a paper to 2010 International Conference on Image Processing.
	Human Pose Estimation for Real-Time Experience Sharing Using Avatars Stanford, 2008.10-2009.2 Proposed an algorithm to estimate human pose in interactive frame rate based on a single webcam and no markers. Designed and implemented a system to automatically monitor user health indices during exercise at home. Published a journal paper on Journal of Entertainment Computing 2009 as the leading author. Published a paper in the Proceedings of the 3rd International Conference on Intelligent Technologies for Interactive Entertainment (INTETAIN 2009).
	IntentSearch Microsoft Research Asia, 2007.7-2008.6 The technology has been shipped to Microsoft Bing. (Try it: Search an image by keyword, hover mouse over an image, then click on "show similar images") Designed and prototyped the system as a real-time large-scale automatic system work directly on top of Microsoft Live Image Search and Google Image Search to re-rank the text-based search results according to user interaction. Designed and implemented an interactive image search, re-rank, organize, browsing interface using Silverlight 2.0 inspired by Seadragon and Microsoft Surface. Dr. Harry Shum, Corporate Vice President of Microsoft, showed IntentSearch as a live demo in his keynote speech in the 17th International World Wide Web Conference (WWW 2008). Published a paper on ACM Multimedia 2008 as the leading author. The live demo show was accepted in the ACM Multimedia 2008 Technical Demonstrations Track.
	Automatic Photo Genre Classification (APGC) Tsinghua University, 2007.5-2008.6 Collaborated with researchers in HP Lab Palo Alto, designed algorithms for automatic image content analysis, scene understanding, keyword annotation, etc.
	EasyAlbum Microsoft Research Asia, 2006.6-2007.6 Designed and prototyped the system as a flexible tool to make use of human perception ability in the process of family album tagging. Integrated features including face detection, face clustering, scene clustering, re-ranking, search, and collage. See project page for details. Collaborated with technical transfer team of Microsoft, released EasyAlbum 1.0. Published a paper on SIGCHI 2007 as the leading author.
	Combining Stroke-based Image Search with Traditional Image Search Tsinghua University, 2007.1-2007.5 Proposed machine learning algorithm to seamlessly combine relevance feedback inside images through drawling strokes with traditional relevance feedback methods. Published a paper on ACM Multimedia 2007 as the leading author.
	Learning Object Cutout from Single Example Microsoft Research Asia, 2006.10-2007.2 Designed algorithms to segment foreground from images with little human interaction. Published a paper on IEEE CVPR 2008 as the leading author.
	Active Learning for Query Design in Image Retrieval Tsinghua University, 2006.9-2006.10 Proposed to query for labels of multiple images in a batch mode in image retrieval. Research results accepted by ICASSP 2008.
	2D Barcode Localization and Recognition Tsinghua University, 2005.11-2006.6 Boosted the speed of anti-screw algorithm by reformulating the algorithm and optimizing the code. Developed and optimized Arbitrary Accuracy Math Component for Symbian, optimized and shipped to mobile phone.
	Traditional Chinese Calligraphic Art Generating System Tsinghua University, 2005.1-2005.6 Implemented traditional Chinese Calligraphic Brush Model which enables composing by touch pad. Took part in designing learning algorithm to automatically generate calligraphic art with personal style.
	Coal Tippler Surveillance and Control System Qinhuangdao Port Group, Co., Ltd., 2004.6-2004.8. Assembled real-time system of Automatic Line of Coal Tipplers, which is still in field use. Enhanced working condition of the port workers greatly by enabling remote surveillance and control.

Publications  

Peer Reviewed Conference Proceedings

1. Jingyu Cui, Fang Wen, Xiaoou Tang, Real Time Google and Live Image Search Re-ranking, in Proceedings of ACM Multimedia, Vancouver, BC, Canada, October, 2008. [Bibtex] [pdf 2.2M]
2. Jingyu Cui, Fang Wen, Xiaoou Tang, IntentSearch: Interactive On-line Image Search Re-ranking, in Proceedings of ACM Multimedia, Vancouver, BC, Canada, October, 2008. [Bibtex] [pdf 2.4M]
3. Jingyu Cui, Qiong Yang, Fang Wen, Qiying Wu, Changshui Zhang, Luc Van Cool, Xiaoou Tang, Transductive Object Cutout, in Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), Anchorage, Alaska, USA, June, 2008. [Bibtex] [pdf 1.2M] [ps 0.8M]
4. Jingyu Cui, Fang Wen, Rong Xiao, Yuandong Tian, Xiaoou Tang, EasyAlbum: An Interactive Photo Annotation System, in Proceedings of ACM SIGCHI Conference on Human Factors in Computing Systems, San Jose, California, USA, 2007. [Bibtex] [pdf 4.4M] [ps 6.2M]
5. Jingyu Cui, Saurabh Mathur, Michele Covell, Vivek Kwatra, Mei Han, Example-based Image Compression, in Proceedings of The International Conference on Image Processing (ICIP), 2010. [Bibtex]
6. Jingyu Cui, Fang Wen, Xiaoou Tang, User Intention Modeling for Interactive Image Retrieval, in Proceedings of IEEE International Conference on Multimedia & Expo, International Workshop on Visual Content Identification and Search (VCIDS), 2010. [Bibtex]
7. Jingyu Cui, Changshui Zhang, Combining Stroke-based and Selection-based Relevance Feedback for Content-based Image Retrieval, in Proceedings of ACM Multimedia, Augsburg, Germany, September, 2007. [Bibtex] [pdf 1.3M] [ps 1.5M]
8. Jingyu Cui, Changshui Zhang, Towards Optimal Query Design for Relevance Feedback in Image Retrieval, in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing, 2008. (ICASSP 2008). [Bibtex] [pdf 0.3M] [ps 0.2M]
9. Yasmin Aghajan, Joyca Lacroix, Jingyu Cui, Aart van Halteren, Hamid Aghajan, Home Exercise in a Social Context: Real-Time Experience Sharing Using Avatars, in Proceedings of The 3rd International Conference on Intelligent Technologies for Interactive Entertainment (INTETAIN 2009).[Bibtex] [pdf 0.7M]
10. Wei Liu, Buyue Qian, Jingyu Cui, Jianzhuang Liu, Spectral Kernel Learning for Semi-Supervised Classification, in Proceedings of Twenty-first International Joint Conference on Artificial Intelligence (IJCAI-09).[Bibtex] [pdf 0.4M]

Journal Articles

1. Jingyu Cui, Guillem Pratx, Sven Prevrhal, Craig S. Levin, Fully 3D List-Mode Time-of-Flight PET Image Reconstruction on GPUs Using CUDA, Medical Physics, vol. 12, no. 12, pp. 6775-6786, 2011.[Bibtex] [pdf 1.1M]
2. Jingyu Cui, Yasmin Aghajan, Joyca Lacroix, Aart van Halteren, Hamid Aghajan, Home Exercise in a Social Context: Real-Time Experience Sharing using Avatars, Journal of Entertainment Computing, in press.[Bibtex] [pdf 1.2M]
3. S. Napel, C.F. Beaulieu, C. Rodriguez, J. Cui, J. Xu, A. Gupta, Danny Korenblum, H. Greenspan, Y. Ma, D.L. Rubin, Automated Retrieval of CT Images of Liver Lesions Based on Image Similarity: Method and Preliminary Results, accepted by Radiology, November 5, 2009.[Bibtex]

Book Chapters

1. Guillem Pratx, Jingyu Cui, Sven Prevrhal, Craig S. Levin, 3-D Tomographic Image Reconstruction from Randomly Ordered Lines with CUDA, GPU Computing Gems, in press.

Conference Abstracts and Presentations

1. Jingyu Cui, Guillem Pratx, Sven Prevrhal, Bin Zhang, Lingxiong Shao, Craig S. Levin, Data-Driven Spatially-Varying Point Spread Function for List-Mode PET Reconstruction on GPU, IEEE 2011 Medical Imaging Conference (IEEE MIC 2011).
2. Jingyu Cui, Garry Chinn, Craig S. Levin, Fast and Accurate 3D Compton Cone Projections on GPU Using CUDA, IEEE 2011 Medical Imaging Conference (IEEE MIC 2011).
3. Jingyu Cui, Sven Prevrhal, Guillem Pratx, Lingxiong Shao, Craig S. Levin, Fully 3-D List-mode Positron Emission Tomography Image Reconstruction on a Multi-GPU Cluster, 11th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine (Fully3D 2011).
4. Jingyu Cui, Guillem Pratx, Sven Prevrhal, Craig S. Levin, Fully 3-D Time-of-Flight Positron Emission Tomography Image Reconstruction from List-mode Data Using GPUs and CUDA, IEEE 2010 Nuclear Science Symposium and Medical Imaging Conference (IEEE NSS/MIC 2010).
5. Jingyu Cui, Guillem Pratx, Sven Prevrhal, LingXiong Shao, Craig S. Levin, Highly Parallel Image Reconstruction for Positron Emission Tomography (PET), 2010 GPU Technology Conference (GTC 2010).
6. S. Napel, C.F. Beaulieu, J. Faruque, C. Rodriguez, D. Korenblum, J. Cui, J. Xu, A. Gupta, H. Greenspan, G. Tye, D.L. Rubin, Content-based Image Retrieval: Portal Venous CT of Liver Lesions, accepted by SIIM 2010.[Bibtex]
7. Sven Prevrhal, Jingyu Cui, Modern architecture for massively parallel medical tomographic image reconstruction on a GPU cluster, 2010 GPU Technology Conference (GTC 2010).
8. Eric González, Jingyu Cui, Guillem Pratx, Matthew Bieniosek, Peter D. Olcott, Craig S. Levin, Point Spread Function for PET Detectors Based on the Probability Density Function of the Line Segment, IEEE 2011 Medical Imaging Conference (IEEE MIC 2011).
9. Virginia Spanoudaki, Jingyu Cui, Jung-Yeol Yeom, Craig S. Levin, Evaluation of the Benefits of Depth of Interaction Compensated Time of Flight PET, 2011 World Molecular Imaging Congress.

Theses

1. M.S. Thesis: Machine Learning for Content Based Image Retrieval, Automation Department, Tsinghua University, June 2008.
2. B.E. Thesis: 3D Calligraphic Brush Model for Realistic Manuscript Rendering, Automation Department, Tsinghua University, June 2005.

Professional Services and Memberships

1. ACM Student Member.
2. IEEE Student Member.
3. SIAM Student Member.
4. Paper reviewer for:
• Medical Physics;
• IEEE Multimedia;
• ACM Transactions on Intelligent Systems and Technology (TIST);
• Neurocomputing;
• SIGGRAPH 2011;
• 2008 International Conference on Intelligent User Interfaces (IUI '08).

Industry

During Oct. 2006 and March 2007, I co-founded CoolBar, co. ltd. as the CTO. I am responsible for the technology evaluation and justification of the company.

Before that, our team of six was the champion of 2006 Maryland Smith China Business Plan Competition, and got $25,000 award as the seed money for CoolBar. We also won the 1st prize in Microsoft Business Plan Competition, with 25,000rmb award in the same year.

Life

I am a big fan of Photography, Ping pong, Pool, Traveling, and keep a close eye on recent progress in start-ups in IT field. Visit my blog to see my photos and thinking. I am also a good runner. I attended half distance (21km) competition of Beijing International Marathon, 2003.

Two places change my life dramatically after high school. One is Tsinghua University, the other is MSRA.

I was lucky to be admitted by Tsinghua after the visit to Hong Kong attending National Mathematical Olympiad 2001. In Tsinghua, I not only received academic training, but also experienced communication with lots of people, and served as tutor of 150 undergraduate students of the '05 grade in department of automation for 2 years (2005.8-2007.8).

Since June 2006, I worked as a visit student In MSRA. I met with lots of senior researchers there, and made many good friends. MSRA not only boosts my research, but also helps me travel more. I went to Palo Alto, San Jose and San Francisco alone to attend CHI 2007 as my first trip to US in May, 2007, and visited Seattle, Redmond, as one of 10 "Best Intern"s selected from Microsoft Research Asia, and invited by Mr. Bill Gates to his house to attend a BBQ in June, 2007.

Awards

• IEEE Medical Imaging Conference Trainee Grant, 2011.
• Pan Wen-Yuan Foundation Scholarship, 2010.
• IEEE Medical Imaging Conference Trainee Grant, 2010.
• Engineering Fellowship, Stanford University, 2008-2009.
• Honor of Outstanding Master Graduate Student, Tsinghua University, 30 over 3000+ Master graduate students, 2008.
• Honor of Outstanding Master Thesis, Tsinghua University, 2008.
• Honored as one of the 10 “Best Interns” in Microsoft Research Asia out of 300 interns, invited by Mr. Bill Gates to visit Microsoft headquarters, have dinner at his house in Seattle, 2007.
• First Prize in Smith China Business Plan Competition, University of Maryland, with $25,000 award, 2006.
• First Prize in Microsoft Business Plan Competition, with rmb25,000 award, 2006.
• University President Scholarship (Highest Honor for students in Tsinghua University), 5 over 10000+ undergraduate students in Tsinghua University, 2005.
• First Class Scholarship in Tsinghua University, 2001, 2002, 2003, 2004.
• Title of Excellent Student Cadre of Tsinghua University, 2003.
• Honor of Outstanding Graduates, Tsinghua University, 50 over 3000+ students, 2005.
• Outstanding graduate student in Beijing City, Beijing Education Committee, 2005.
• All-round Excellent Student of Beijing, 2005.

References

Miscellaneous

• Blog:
• Legacy MSN Space
• WordPress Space
• Travel and Photography:
• Picasa Web Album
• Cooliris 3D Wall of My Albums
• Profiles: LinkedIn
• Programming: I am fond of implementing ideas to see how things work out in real applications, and also interested in build graphical user interfaces (GUI) to facilitate human computer interaction. C/C++, C# and Java are most used tools. I can also program with Visual Basic, ASP.NET, HTML/CSS/Javascript, PHP, Actionscript, WPF, and Silverlight. I've experienced several projects related to SQL database programming, and have implemented an application oriented file database myself with high performance (trade off from limited functionality). As for academic programming, I am proficient in scripts such as Matlab, Scilab, and Octave, which I use in laboratories to try out small algorithms.

Links

• Conferences:  CVPR 2012, ECCV 2012, ISBI 2012, SIGGRAPH 2011, ICCV 2011, ICML 2011, CVPR 2011, ICCP 2011, I3D 2011, MICCAI 2011, SC 2011, IEEE MIC 2011, WMIC 2011, Fully 3D 2011, SNM 2011, SPIE Medical Imaging 2011, ISBI 2011, 2011 International CES, SIGGRAPH 2010, SIGGRAPHASIA 2010, IEEE MIC 2010, Google I/O 2010, SC 2010, GTC 2010, ICASSP 2010, CVPR 2010, ECCV 2010, ICIP 2010, SC 2009, SIGGRAPH 2009, CVPR 2009, ICCV 2009, ICML 2009, IJCAI 2009, CVPR 2008, UIST 2008, ECCV 2008, ACM Multimedia 2008.
• Labs: Bigeye, MSRA VC Group, Google Research, HP Lab Palo Alto.

Last updated: June 28, 2011.