Background extraction is generally the first step for many computer vision and augmented reality applications. Most existing methods, which assume a clean background shot, are not suitable for video sequences such as highway traffic videos containing complex foreground movements which may not have a clean background shot. Therefore, this projects focuses on developing algorithms for robust background extraction.
Motion capture is an important technique for animation. However, it requires expensive equipment to achieve precision. This project focuses on capturing the human motion suing simple and cheap depth sensors. Furthermore, the motion of camera can be used to manipulate a virtual camera for camera planning.
Title:Non-photorealistic Chinese Painting Animation
Description:
Chinese ink paintings are pervasive in Asia and some art departments in Europe, and the Americas are offering classes to teach this unique style of paintings. Historically, scholars with accomplishments in politics, literature and science learned this art, and their products depict beauty and harmony of nature through brush strokes that shape and move objects such as water. In still paintings, special techniques are employed to deliver the endless spirit and flowing sense of water, yet dynamic water movement can enhance the effect dramatically. There have been several recent developments to augment traditional static paintings with moving objects. For example, the movie titled "Moving of the Riverside Scene at Chingming Festival" received positive reviews when shown at WorldExpo 2010 (Shanghai, China). In the movie, animated characters and dynamic objects (e.g., rivers) are seamlessly blended into famous classical Chinese paintings. Those animations bring the original static painting to life. This project focuses on how to animate elements in the Chinese paintings.
Stereoscopic 3D (S3D) movies and animations have become more and more popular after the appearance of Avatar. However, it is a necessary but challenging task for creative producers to have an idea how the target audience might perceive when watching a stereoscopic film in a cinema during production and setting stereoscopic parameter is a time-consuming job and relies on animator's experience. This project focuses on providing a visualization tool to provide indication of possible stereoscopic perception and to accelerate the S3D parameter setting process.
Stereoscopic 3D animation become popular after the enter of Avatar. However,
Ultra-high-definition displays are available on the market, and vector graphics are possible solutions to describe various multimedia contents in versatile applications because of its compactness and scalability. Therefore, this project focuses on proposing a new data structure and vector algorithm to preserve complex details of arbitrary images and maintain interactive editing and manipulation abilities.
Furthermore, high resolution textures are determinant for high rendering quality in gaming and movie industries, but burdens in memory usage, data transmission bandwidth, and rendering efficiency. Therefore, it is desirable to shade 3D objects with vector images such as Scalable Vector Graphics (SVG) for compactness and resolution independence. We also focus on applying the vector image on rendering 3D objects.
Comics and Manga are popular medium for conveying ideas and entertainment. Manga, Japanese comics, have gained worldwide audience from not only Asia but also North America and Europe. With advance in mobile technologies, Manga digitization is important for distribution on these devices. However, scanned conversion still dominates digitization and requires manual intervention to clear up noise, artifacts, and distortions. The process is strenuous and time-consuming, and results are incapable to transform to different resolutions after scanned. Furthermore, the size of a page cannot be fitted into the mobile devices and thus, we would like to segment them for better readability
Image processing has been research topics for decades. Therefore, our group has focuses on extracting information from photos. We also consider how to manipulate image size in single view and dual views.
A picture say a thousand words. Human is good at interpreting data from visualizing a set of numbers instead of number itself. In this project we try to help geologists develop a system to visualize the data collected.
Group motion is important for animation and game. However, the computation is time consuming. Therefore, we try to explore the possiblity of presimulating the motion of groups and then dynamically synthesize new motions from existing clips.
Global illumination is important for many applications such as animation, computer games, architecture, design and etc. However, the main issue in this field is the slow converging rate of algorithms. This research focus on improving the efficiency.
