Abstract: A method for robust human face tracking in the presence of multiple facial images includes taking a frame from a color video sequence as a current input image; filtering the current input image to form a filtered image; estimating the locations and sizes of faces in the filtered image based on a projection histogram of the filtered image; estimating face motion in the filtered image; and outputting the location and size of the tracked faces within the filtered image.

Abstract: A method and apparatus for tracking of objects in video images. The tracking process includes the steps of: receiving an initial image with the object to be tracked; performing a color histogram on the initial image in the region where the object is located; receiving the next frame of the video sequence and specifying candidate regions in that image; using color histogram analysis on the candidate regions to find a match performing motion analysis to correct the analysis and to provide a prediction for the next frame; and, identifying and providing the tracked object's image for output. Apparatus that can use this method can be either video capture or video display. In video capture, a user interface allows the user to designate the object to be tracked during capture. The tracking module than controls the capture system to track that object. In video display systems, a user interface allows the user to identify the object in the display that should be tracked.

Abstract: A hierarchical system for object-based audiovisual descriptive tagging of images for information retrieval, editing, and manipulation, includes: an object-based selection mechanism for selecting an object of interest in said image; hierarchical data structure generation means for generating a hierarchical data structure for said image and for associating auxiliary information with said image; and a transmission/storage mechanism for storing the image and the hierarchical data structure. A hierarchical method for object-based audiovisual descriptive tagging of images for information retrieval, editing, and manipulation, includes: selecting an object of interest in said image with an object-based selection mechanism; generating a hierarchical data structure for said image and for associating auxiliary information with said image; and transmitting/storing the image and the hierarchical data structure.

Abstract: A method for content-scalable shape representation is described. In content-scalable shape representation, the more important parts of the image are encoded using a larger number of bits within a single hierarchical layer, or within one or more layers of multiple hierarchical layers. This representation may be used with the framework of a hierarchical vertex-based shape representation technique. The method described herein provides for the representation of shapes either by selected verticles along a contour, or by binary bit maps. The resulting representation facilitates content-scalable coding.

Abstract: A method for coding multiple layer, vertex-based shape representation and bit stream scalable shape includes selecting vertices in a hierarchical scheme; defining multiple layers of vertices; arranging M layers of vertices in order, from the most salient layer to the least salient layer; placing vertices in their natural order along their image contour within each layer; storing vertices from each layer separately; and recording vertex order information for each vertex in each layer of vertices. Four variations of the method for multiple layer vertex representation and coding are disclosed where the correct order information is coded for correct reconstruction of the shape information.

Abstract: A method of vertex-based hierarchical shape representation and coding includes selecting an image of interest, defining a boundary about an object of interest within the image of interest, wherein the boundary is defined by a set of contour points, selecting a number of layers, each layer having a predefined saliency parameter associated therewith, wherein the saliency parameter defines a range of layers from a most salient layer to a least salient layer, identifying the most salient contour points in the boundary and placing the most salient contour points in the most salient layer, identifying the next most salient contour points in the boundary and placing the next most salient contour points in the next most salient layer until all of the selected layers have contour points placed therein, and hierarchically storing the layers in a database.