Abstract

Topic 1: Parallel Computation for Image Processing and Data Visualization

This talk is comprised of four parts that include, GPU programming, Natural User Interface, 3D Reconstruction and Future challenges. Recently, with the rapid development of graphics cards, a graphics processing unit (GPU) with a highly parallel many-core architecture has been widely used in 2D image processing, 3D data analysis, visualization, and other fields. Natural user interfaces (NUIs) provide human computer interaction (HCI) with natural and intuitive operation interfaces, such as using human gestures and voice. We have developed a real-time NUI engine using a web camera, which implements real-time image processing using GPU programming. During mobile robot navigation, more than 100,000 laser points and 20 images are sensed in large-scale environment. Using GPU programming, the large-scale terrain photorealistic visualization system achieves high-speed and high quality performances. There is a new hope of achieving "Interactive Data Mining and Visualization (IDMV)". As quoted by Edward Tufte, "Every single pixel should testify directly to content". We aim to accelerate the traditional CPU-based data mining algorithm by GPU programming technology. The data mining results are represented with intuitive and understandable graph. The users have NUI interface to interact the mining processing.

Topic 2: Real-time Camera Natural User Interface Engine

Natural User Interfaces (NUIs) are a novel way to provide Human Computer Interaction (HCI) with natural and intuitive operation interfaces, such as using human gestures and voice. This talk presents a real-time NUI engine architecture using a web camera as a means of implementing NUI applications with an inexpensive device. The engine integrates the OpenCV library, the CUDA toolkit, and the DirectX SDK. The OpenCV library is used to capture video via a simple web camera, implementing real-time image processing using Graphic Processing Unit (GPU) programming, and it presents the NUI applications using the DirectX SDK; for example, to implant a 3D object in a captured scene and play sounds. To verify the efficacy of our proposed NUI engine, we utilized it in the development and implementation of several mixed reality games and touch-less operation applications. Our results confirm that the methods of the engine are implemented in real time and the interactive operations are intuitive.