ISRG Schedule, Fall 2007

The AMD Graphics Product Group Developer Tools Group is currently designing software to help graphics developers in two areas: identifying and correcting performance bottlenecks and accelerating developers' time to market in the handheld world. While the aggressive releases of new hardware for both desktop and handheld devices are resolving some graphics and performance issues, there are still many preprocessing steps that can be performed to get the most out of the hardware. Additionally, new API and hardware features need to be exposed and fully supported so that developers can provide graphics enthusiasts with the ultimate visual experience.

Body bending and body twisting are kinds of non-rigid and non-periodic patient motion which may occur during SPECT imaging. We develop a body deformation correction method to improve the 3D reconstruction of SPECT imaging. In this paper, we create a body deformation model, measure the body deformation and improve the reconstruction for the body volume. The deformation model proposed here has three degrees of freedom: twisting angle, bending direction and bending angle. Based on such a model, we estimate these deformation parameters by a set of skin-based marker’s coordinates. We also design an experiment to show that our body model works well for real patients. Then a modified MLEM algorithm is proposed to adopt the deformation correction into the reconstruction. Finally, we create a set of reconstruction results with or without deformation correction by the NCAT phantom. Such simulation results are compared to show that our method can benefit the MLEM 3D reconstruction method.

Keywords: Bending, Twisting, Deformation, Motion Correction

More and more researchers are focusing on the management, querying and pattern mining of streaming data. The visualization of streaming data, however, is still a very new topic. In this proposal, we discuss our plan to construct a multivariate streaming data visualization system. Three subtasks are identified, including streaming data abstraction, visualization and interaction techniques for streaming data, and visualizing change in data streams. An overview of proposed solutions is provided.

In a matter of just few years, the programmable graphics processor unit has evolved into an absolute computing workhouse. In 2006, compared with around 30GFLOPS by Intel Core2 Duo 3.0Hz, G80GL archives up to 350FLOPS. With multiple cores driven by very high memory bandwidth, todays GPUs offer incredible resources for both graphics and non-graphics processing. CUDA (Compute Unified Device Architecture) is a new hardware and software architecture for issuing and managing computations on the GPU as a data-parallel computing device without the need of mapping them to a graphics API. It provides a massively multithreaded architecture with up to 128 processor cores and thousands of threads in programmable in C and capable of hundreds of billions of floating-point operations each second. In this talk, I will describe the architecture and programming model of CUDA, give some examples that illustrate the data-parallel algorithm, present my summer intern work in CUDA area, and close with the demo.

Nowadays, many first-person virtual reality and 3D game applications employ the use of camera oscillations to give the user a more realistic sensation of walking or running. However, the oscillation that is applied to the camera varies from application to application. The HIVE research group is currently developing a tool whose purpose is to study each one of these different oscillatory movements and evaluate their effectiveness in increasing user's immersion in a virtual environment (VE). In addition, the application also provides users with some extra help in navigating through the VE by directing their attention to important objects that should paid attention to or places that should be visited. This presentation will explain how the current version of the tool works and what is going to be enhanced before the users studies are carried out.

Visual analytics is an emerging field which uses visual techniques to interact with users in the analytical reasoning process. Users can choose the most appropriate representation that conveys the important content of their data by acting upon different visual displays. The data itself has many features of interest, including clusters, trends (commonalities) and anomalies. Most visualization techniques currently focus on the discovery of trends and other relations, where uncommon phenomena are treated as outliers and are either removed from the datasets or de-emphasized on the visual displays. Much less work has been done on the visual analysis of outliers, or anomalies. In this thesis, I introduced a method to identify the different levels of "outlierness" by using interactive selection and I compared two different methods of interaction while investigating the anomalies in the dataset. In one approach, the outlier degree was treated as a dimension of the dataset. The other approach created a separate dialog box for anomaly interaction which includes changing parameters and updating the graph. I will also do user evaluation to compare these two methods. The application was developed based on the XMDV tool at WPI.

Dimensional Stacking is a technique for displaying multivariate data in two dimensional screen space. This technique involves the discretization and recursive embedding of dimensions, each resulting N-dimensional bin occupying a unique position on the screen. This paper describes the extension of this technique to a three dimensional projection. In addition to the visual enhancements, hashing was used to improve the scalability of records and dimensions. The resulting visualization was evaluated by a usability study.