|
Lin ChenSoftware Engineer, Google
|
Short Biography
I am a software engineer at Google, working on Android Camera Machine Intellegence. Before Google I was a software engineer in Snap Inc. working on creative camera.
I got my PhD at Department of Computer Science and Engineering, Arizona State University, under the supervision of Prof. Baoxin Li. My research interest includes computer vision & Graphics, applied machine learning and data mining. Before coming to ASU, I received my B.S. and M.S. degree in 2008 and 2011 from the computer science and technology department, Shandong University.
Selected Publications
Conferences
- Embedded Supervised Feature Selection for Multi-class Data.
Lin Chen, Jiliang Tang, Baoxin Li.
Proceedings of the 2017 SIAM International Conference on Data Mining June, 2017. - Clustering-based Joint Feature Selection for Semantic Attribute Prediction.
Lin Chen, Baoxin Li.
International Joint Conference on Artificial Intelligence (IJCAI) June, 2016. - Fusing Pointwise and Pairwise Labels for Supporting Personalized Image Retrieval.
Lin Chen, Peng Zhang, Baoxin Li.
ACM International Conference on Multimedia Retrieval (ICMR) June, 2015. [Paper] - Instructive Video Retrieval For Surgical Skill Coaching Using Attribute Learning.
Lin Chen, Qiang Zhang, Peng Zhang and Baoxin Li.
IEEE International Conference on Multimedia and Expo (ICME) June, 2015. [Paper] - Instructive Video Retrieval Based on Hybrid Ranking and Attribute Learning.
Lin Chen, Peng Zhang and Baoxin Li.
ACM Multimedia (ACMMM) November 2014. [Paper] - Predicting Attributes via Relative Multi-task Learning.
Lin Chen, Qiang Zhang and Baoxin Li.
Computer Vision and Pattern Recognition (CVPR) 2014. [Paper] - Learning Skill-Defining Latent Space in Video-Based Analysis of Surgical Expertise - A Multi-Stream Fusion Approach.
Lin Chen, Qiang Zhang, Qiongjie Tian and Baoxin Li.
Medicine Meets Virtual Reality (NextMed / MMVR) 2013. [Paper] - Enhancing Fundamentals of Laparoscopic Surgery Trainer Box via Designing A Multi-Sensor Feedback System.
Qiongjie Tian*, Lin Chen*, Qiang Zhang and Baoxin Li. (* Equally contributed)
Medicine Meets Virtual Reality (NextMed / MMVR) 2013. [Paper] - Video-based analysis of motion skills in simulation-based surgical training.
Qiang Zhang, Lin Chen, Qiongjie Tian and Baoxin Li.
IS&T/SPIE Electronic Imaging 2014. [Paper] - Anisotropic Resizing of Model with Geometric Textures.
Lin Chen and Xiangxu Meng
SIAM/ACM Joint Conference on Geometric and Physical Modeling (SPM) 2009. [Paper]
Journals
- User-adaptive image retrieval via fusing pointwise and pairwise labels.
Lin Chen, Peng Zhang, Baoxin Li.
International Journal of Multimedia Information Retrieval Dec., 2015. [Paper] - Max-margin Multi-attribute Learning with Low-rank Constraint.
Qiang Zhang, Lin Chen and Baoxin Li.
IEEE Transactions on Image Processing, Vol. 23 2014. [Paper] - Anisotropic Resizing and Deformation Preserving Geometric Texture.
Lin Chen and Xiangxu Meng.
Science in China Series F-Information Sciences Information Sciences 2010. [Paper]
Selected Projects
| Context based image / video filter We train a faster R-CNN model to analyze the context of the images / videos, e.g., detect and recognize visual objects, and add interesting filters /effects according to the context, e.g., change the sunset sky to starry night sky. |
| Custom stickers We build a tool based on Grab-cut algorithm which is able to extract the foreground of the users' interested area. The extracted foreground images can then be used as stickers to decorate new image / videos. |
| Context based image search We build a search system which allows the user to search their snaps based on context, e.g., find photos / videos with bicycles. |
| Pixelation Brush We build an interactive drawing tools which allows the user to draw special effects on image / videos real time with the help of fast GPU rendering. A patent is filed from this project. |
| View recommendation for 360-degree panorama image I worked on this project during my 2015 summer research intern in Nokia Technologies. We are working on the problem that how to display the 360 degree panorama image in traditional display devices while at the same time preserving aesthetic visual quality. |
| Image ranking/classification based on semantic attributes Images can be described by semantic attributes, e.g., natural, open, etc., which are able to brige the low-level image representation and high-level human understanding. We proposed a set of machine learning algorithm to improve the image classification/ranking performance based on semantic attributes. |
| Skill evaluation on video-based surgical training video-based surgical training system has been well adopted to train new surgeons. On this project, we developed a system to analyze surgeons' skill and recommend tutorial video for self-skill improvement. |
| Geometric texture transfer on model deformation Geometric texture is usually distored during 3D model deformation, especially when resizing operation invloved. In this project, we implement an algorithm which can keep the appearance of geometric textures during model deformation based on geometric synthesis. |
Selected Awards
| 2015 | University Graduate Fellowship |
| 2014 | SACMMM'14 student travel award |
| 2010 | IBM Chinese Excellent Student Scholarship |
| 2008 | Outstanding students of Shandong Province |
| 2007 | MCM(The Mathematical Contest in Modeling)/ICM(The Interdisciplinary Contest in Modeling) Honorable mention |
| 2007 | Chinese national scholarship for outstanding students |
| 2007 | HP Scholarship for Outstanding Chinese Students |
| 2006 | Chinese Undergraduate Mathematical Contest in Modeling (CUMCM) National second prize |
Service
| Teaching Assistant: CSE 310 Data Structure and Algorithm, CSE 591 Fundation of Algorithm, Arizona State University; Computer Graphics, Shandong University |
| External Reviewer: CVPR, ICCV, CIKM, IJCAI |
| Student Volunteer: CVPR'14, ACMMM'11 |