Vanyi Apr 2022 - Present
After spending 3 years at Kujiale, I eventually chose to leave and then joined Wanyi Technology. During my three years at Kujiale, my technical and managerial skills have improved significantly. The EGS rendering engine developed by our team has achieved a rendering performance improvement of 3-4 times. I have also been promoted from a developer to a senior technical expert, leading a research and development team of up to 10 people. I have met many outstanding colleagues here and have been appreciated by the company. I have chosen to leave now mainly because I hope my rendering technology can be exercised on more platforms, not just on the web. I am very grateful for the experience at Kujiale, which has allowed me to grow a lot.
In vanyi I still focus on real-time rendering. We created a engine called VRE, click here to see more.
In vanyi I still focus on real-time rendering. We created a engine called VRE, click here to see more.
Kujiale Apr 2019 - Mar 2022
Finally I left model at on Mar 2019, it was a very good experience at the beginning 2 years. I have to say that I learned so much from the CTO and other co-workers. It was such a dream team. But I have to move on I suppose. Since Apr 1st, I got the new job from Kujiale, an online SaaS-enabled home decoration and design platform. I finally had got my own team and work on the 3D engine optimization work.
MODELO website Dec 2016- Mar 2019
I entered into the Modelo team since the ending of 2016 to help the front end rendering part work. the website is http://modelo.io
It is a real-time 3D design collaboration, showcasing and project management platform. You can check the website to see how cool it is.
Again it's a pure 3D job, all my passion is split into 2D and 3D work. I love them both!
It is a real-time 3D design collaboration, showcasing and project management platform. You can check the website to see how cool it is.
Again it's a pure 3D job, all my passion is split into 2D and 3D work. I love them both!
DSST Object Tracking Sep 2016 - Dec 2016
I reference and implement the algorithm of paper “Accurate Scale Estimation for Robust Visual Tracking”, also apply the algorithm on mobile device with C code library and API
Click here to see more.
Click here to see more.
Joint cascade face detection and alignment July 2015 - May 2016
This is the project in Tuya, it used a very smart algorithm to get the face area and key points at same time during a very short time.
The detection result is quit good.
Click here to see more.
The detection result is quit good.
Click here to see more.
Face Recognition with DeepID Jan 2015 - May 2015
Since Deep Learning is so hot now, we are trying to implement the DeepID algorithm for high accuracy of Face Recognition.
Click here to see more.
Click here to see more.
Multi-View 3D Object Reconstruction 3/20/2014 - 10/5/2014
Reconstruct an object into a 3D model with several pictures with different views. Use SIFT algorithm to extract the image feature points and match them with each other. Calculate the camera matrix and the 3D coordinates. Reform the point cloud into mesh with the images as texture mapping to get the final 3D model
Click here to see more.
Click here to see more.
Face Detection and Face Recognition 6/15/2014-12/15/2014
Apply the Haar classifier to get the face area from images and videos, then train the faces with Fisher algorithm. The accuracy for the detection and recognition are 96% and 85%. The next step is to apply the Deep Learning algorithm into the face recognition.
Click here to see more.
Click here to see more.
Object Tracking On Cellphone 10/15/2013-2/20/2014
A online open source algorithm which is used for object tracking. It learns from the movement of the object and detects the object even if it's blocked. Moved the project from PC to cellphone.
Click here to see more.
Click here to see more.
CUDA Video Checking 11/1/2013-25/3/2014
Using GPU cuda to check if the video is ok to release. It's originally with a check person while now is using the computer. With CUDA acceleration, we aim to get at leate 8 times faster. We have currently black screen, green screen, black edge, mosiac , and static frame checking. Since it's still working, the demo video will come soon!
AppShare 6/15/2013-9/20/2013
It's a project cooperate with the Japanese Client-NTT Docomo, we offered a framework to use wifi-direct to share apps based on one authorized host to several clients devices and delete the app automatically after the connection is off.
Click here to see more.
Fast animation with 3D model 5/5/2013-9/1/2013
This is a project I did during my free time. I used kinect and Maya to drive a model to make animation. The frame rate is 30 FPS.
Totally meet the requirement of normal animation. Looks very cool.
Click here to see more.
Totally meet the requirement of normal animation. Looks very cool.
Click here to see more.
New Oil Painting Effect 9/20/2012-12/20/2012
Currently I am working on a new filter effect about oil painting. I know there a lot of people tried to do this before. Even Photoshop can do this with a few steps of operation. But I believe what I did is way much better.
Click here to see more.
Click here to see more.
Beyond The Stage January-May 2012
We use interactive media to bring the magic of theater to the audience online, hoping to draw that audience to the theater once again.
This semester we worked in association with Kristoffer Diaz, Lynn Nottage, and the Black Women Playwrights’ Group on two projects: Vera Stark and Chad Deity.
Click here to see more.
This semester we worked in association with Kristoffer Diaz, Lynn Nottage, and the Black Women Playwrights’ Group on two projects: Vera Stark and Chad Deity.
Click here to see more.
Oceanus 3 8/26/2011-11/12/2012
Team Oceanus 3 is continuing the Entertainment Technology Center’s development of two interactive crowd experiences for the 2012 World Expo in Yeosu Korea. All guests entering the 2012 expo will pass through the Expo Digital Gallery, walking under a massive LED ceiling display showing our content. Hundreds of guests at a time will be able to interact with the ETC’s Safe Spot and Sumi Twitter experiences. Safe Spot is an augmented reality experience that uses a network camera to virtually transport people onto the melting sea ice of the arctic and Sumi Twitter is an animated visualization of a Twitter feed that breathes life into tweets with the help of colorful ocean life.
Click here to see more.
Click here to see more.
Hand Animation 5/1/2011-7/25/2011
This is the first paid job I get since I came to America. I worked with a post PHD named Junggon and Professor Nancy Pollard. The project is about creating a hand model with the Jounggon's physical library to create a real time movement. The final purpose is to grasp objects in the virtual world and calculating the force real time. My job is to use the Vuzix company's HMD device and the cyberglove to make the user to be able to wondering in the virtual world plus the mocap system to decide the rotation and the position, which give me a change to get to know three new devices at the same time and combine them all in the same projcet. Using different company's sdk and compile them smoothly is the biggest challenge I met now. There should be some thing more interesting happen. :-)
Click here to see more.
Click here to see more.
Minspeak 1/13/2011-5/1/2011
This is the first big project I do in the ETC, and it's a 15 weeks' project.For this project, we will work with SCS to develop a Minspeak® system designed for teaching English and Spanish as foreign languages. SCS envisions computerized and/or paper-based lessons focusing on the syntax and semantics of a target language. An application could accompany the lessons exploiting Minspeak®'s ability to use motor planning, visual memory, and auditory feedback to teach foreign languages in an interactive, game-like way. Complex language patterns such as the English personal pronoun system and its link to the modal/auxiliary system can take non-English speakers from a state of deep mystification to a clear and memorable grid they can easily learn.
The challenge in this project is for us to develop a prototype that demonstrates Minspeak®’s power as a language teaching and learning system.
Currently, we still just this project. We'll have more details in the following days.
Click here to see more.
The challenge in this project is for us to develop a prototype that demonstrates Minspeak®’s power as a language teaching and learning system.
Currently, we still just this project. We'll have more details in the following days.
Click here to see more.