Jeong-Jun Song

Shape-based 3D model retrieval

Two feature extraction methods for shape similarity based retrieval of 3D object models are presented. The proposed methods, which result in more effective and robust techniques for searching 3D models by similarity, support two essential query modes, namely, query by 3D model and query by 2D image. Our feature extraction scheme is inspired by observation of human behavior in recognizing 3D objects in practice. The process of extracting spatial arrangement from a 3D object surface and 2D shape features from projection images are achieved by adopting curvature distribution of the model surfaces and Fourier descriptors of the projection images.

A comprehensive curriculum for IT education and workforce development: an engineering approach

Noting the shortage of IT professionals nationally [1], we propose a comprehensive curriculum that supports a variety of programs geared to all ages from early school years to retirement and beyond. Current IT workforce development efforts are limited to training, and have not as yet focused on education and professional development. Largely, this is due to a lack of a science underpinning for IT related curricula. Without such a unified science component, a structured organization of information related concepts cannot be derived.Our proposal includes the development of a number of programs addressing the needs of a variety of learners ranging from elementary school through college and beyond. Seven programs, each with a specific emphasis for various groups, are being developed. Such essential issues as industrial-academic liaisons, workforce (re)training, promotional and awareness programs, teacher training, and IT professional role redefinition, are integral pieces of this project. All developments will be firmly founded on the scientific framework of information science and engineering [2].This work is supported by NSF grant DUE-9950168.

3D OBJECT FEATURE EXTRACTION BASED ON SHAPE SIMILARITY

We introduce two complementary feature extraction methods for shape similarity based retrieval of 3D object models. The proposed methods lead us to achieve effectiveness and robustness in searching similar 3D models, and eventually support two essential query modes, namely, query by 3D model and query by 2D image. Our feature extraction scheme is inspired by the observation of human behavior in recognizing 3D objects. The process of extracting spatial arrangement from a 3D object can be considered as using human tactile sensation without visual information. On the other hand, the process of extracting 2D features from multiple views can be considered as examining an object by moving the viewing points (or camera positions). We propose a hybrid method of 3D model identification by object-centered feature extraction, which utilizes the Extended Gaussian Image (EGI) surface normal distribution and distance distributions between object surface points and origin. Another technique need in parallel is a hybrid ...

Analyzing blood cell image to distinguish its abnormalities (poster session)

In this paper, we show the blood-cell image classification system to be able to analyze and distinguish blood cells in the peripheral blood image. To distinguish their abnormalities, we segment red and white-blood cell in an image acquired from microscope with CCD camera and then, apply the various feature extraction algorithms to classify them. In addition to, we use neural network model to reduce multi-variate feature number based on PCA(Principal Component Analysis) to make classifier more efficient. Finally we show that our system has a good experimental result and can be applied to build an aiding system for pathologist.

Polygon-based bounding volume as a spatiotemporal data model for video content access

Indexing, retrieval and delivery of visual and spatio-temporal properties of video objects requires efficient data models and sound operations on the model are mandatory. However, most object-based video data models address only a single aspect of those properties. In this paper, we present an efficient video object representation method that captures the visual, spatial and temporal properties of objects in a video in the form of a unified abstracted data type. The proposed data type is a polygon mesh, named video object mesh, which is defined in a spatio-temporal domain. Based on the application needs, a contour of an object is modeled with a polygonal contour. With the contour and color information of the object, content-based triangularization is performed. A video object in a frame is modeled with two dimensional-polygon mesh. Each vertex in the mesh, color information is embedded for further use. By using motion analysis, a corresponding vertex in the adjacent frame is identified connected to the vertex that is being analyzed. These processes are continued until a video object disappears. The result of these processes is a three dimensional polygon mesh hat models location variant motion and location invariant motion that can not be captured by traditional trajectory based motion model. The proposed model is also useful camera motion analysis. Since a surface shape of a video object mesh has partial information of camera motion.

3D Object Retrieval by Shape Similarity

We introduce a method for shape similarity based retrieval in 3D object model database. The proposed method leads us to achieve effectiveness and robustness in similar 3D object search supporting both query by 3D model and query by 2D image. Our feature extraction mechanism is based on observation of human behavior in recognizing objects. Our process of extracting spatial arrangement of a 3D object by surface point distribution can be considered as using human tactile sensation without visual information. On the other hand, the process of extracting 2D features from multiple views can be considered as examining an object by moving viewpoints(camera positions). We propose shape signatures for 3D object model by measuring features of surface point and the shape distance distribution from multiple views of 3D model. Our method can be directly applied to industrial part retrieval and inspection system where different geometric representations are used.

Towards Retrieval of Visual Information Based on the Semantic Models

Most users want to find visual information based on the semantics of visual contents such as a name of person and an action happening in a scene. However, techniques for content-based image or video retrieval are not mature enough to recognize visual semantic completely. This paper concerns the problem of automated visual content classification that allows semantic exploration of the visual information. To enable semantic based image or visual object retrieval, we propose a new image representation scheme called visual context descriptor (VCD) that is a multidimensional vector in which each element represents the frequency of a unique visual property of an image or a region. VCD utilizes the predetermined quality dimensions (i.e., types of features and quantization levels) and semantic model templates mined in priori. Techniques for creating symbolic representation (called visual term) of visual content and semantic model profile mining and matching have also been explored. The proposed model classification technique utilizes contextual relevance of a visual term to a target semantic class in visual object discrimination. Contextual relevance of a visual cue to a semantic class is determined by using correlation analysis of ground truth samples.