Abstract: A color image processing method is provided. The color image processing method includes the step of: (a) indexing a color image by assigning representative colors of an image to a color space divided into a plurality of regions. The color image processing method may be applied to object-based image processing such as MPEG-7, and fast search and retrieval of multimedia content can be made.

Abstract: A color image processing method for retrieving a color feature descriptor for describing color features of an image is provided. The color image processing method includes the steps of (a) obtaining color vectors of an input image, (b) classifying the color vectors to obtain dominant colors of the input image and the ratios thereof, and (c) representing the dominant colors and the ratios thereof as a color feature descriptor of the input image. The color image processing method is applied to an object-based image processing method, thereby allowing fast search and retrieval of multi-media contents.

Abstract: A database building method for multimedia contents is provided. The database building method for multimedia contents has the steps of (a) accessing an arbitrary site providing multimedia contents through a telecommunications network; (b) calling multimedia contents in by spidering the site; and (c) classifying the multimedia contents data according to the stored addresses and storing them in a predetermined database. Using category information on the corresponding sites, the database building method for multimedia contents according to the present invention semantically classifies multimedia contents and stores them in the corresponding databases. In the database built by the database building method for multimedia contents according to the present invention, multimedia contents which are dispersed on the WWW are well collected and, using category information or URL information, are semantically well classified.

Abstract: A method for retrieving an image texture descriptor for describing texture features of an image, including the steps of (a) filtering input images using predetermined filters having different orientation coefficients, (b) projecting the filtered images onto axes of each predetermined direction to obtain data groups consisting of averages of each directional pixel values, (c) selecting candidate data groups among the data groups by a predetermined classification method, (d) determining a plurality of indicators based on orientation coefficients of the filters used in filtering the candidate data groups, and (e) determining the plurality of indicators as the texture descriptor of the image. The texture descriptors which allow kinds of texture structure present in an image to be perceptually captured can be retrieved.

Abstract: An image retrieval method and apparatus independent of an illumination change are provided. The image retrieval method involves: inputting a query image; detecting an illumination color from the query image and converting the illumination color into a standard illumination color; extracting color information of the query image by using color descriptors; and retrieving a similar image by comparing the extracted color information with color information of a database which converts a variety of images into images of standard illumination colors and extracts and stores the color information of the images in advance. According to the method, without changing the structures of color descriptors or a similarity comparison using color information, an image retrieval independent of illumination changes is enabled by adding only a preceding step of standard illumination conversion.

Abstract: A digital image segmentation method and device is provided. The method includes the steps of grouping an image stream into a plurality of continuous frame groups, setting one arbitrary frame as an I frame and the others as P frames in each group, segmenting the I frame into a plurality of regions, and segmenting the P frames by obtaining motion vectors, using motion prediction, of the segmented I frame regions, respectively, with respect to each of the P frames, by moving the segmented I frame regions by the respective motion vectors, and by setting the moved I frame regions as segmented P frame regions. Accordingly, the segmentation of the I frame is applied to the P frame segmentation, and a conventional segmentation process consuming excessive time with respect to all frames does not need to be performed. Thus, rapid segmentation can be performed.

Abstract: A line based image matching method in which a model image indexed by similar shape descriptors to a query image is retrieved from an image database indexed by line based shape descriptors. The line based image matching method involves: collecting line information of a query image and model images; defining the binary relation between the lines of the query image and the lines of the model images; measuring the compatibility coefficients of the node-label pairs of the query and model images based on the binary relation; and measuring the similarity between the query and model images on the basis of continuous relaxation labeling using the compatibility coefficient.

Abstract: An apparatus for analyzing image texture information and a method therefor are provided. The apparatus for analyzing the image texture information includes a filtering unit for filtering a still image including a plurality of pixels of M rows×N columns with filters having different filtering coefficients, an X axis projector for calculating a gray level mean value of a row of N pixels, for each row, with respect to the filtered plurality of images, and a Y axis projector for calculating a gray level mean value of a column of M pixels, for each column, with respect to the filtered plurality of images. It is possible to describe the orientation and the periodicity of the texture in extracting the texture information.

Abstract: A database building method for multimedia contents is provided. The database building method for multimedia contents has the steps of (a) accessing an arbitrary site providing multimedia contents through a telecommunications network; (b) calling multimedia contents in by spidering the site; and (c) classifying the multimedia contents data according to the stored addresses and storing them in a predetermined database. Using category information on the corresponding sites, the database building method for multimedia contents according to the present invention semantically classifies multimedia contents and stores them in the corresponding databases. In the database built by the database building method for multimedia contents according to the present invention, multimedia contents which are dispersed on the WWW are well collected and, using category information or URL information, are semantically well classified.

Abstract: A method of indexing a high-dimensional vector space, along with a method of quickly retrieving a feature vector having features similar to a query vector from the vector space indexed by the indexing method, are provided. The method of indexing a feature vector space includes the steps of (a) partitioning the feature vector space into a plurality of approximation regions; (b) selecting an arbitrary approximation region to determine whether the selected approximation region is heavily or sparsely distributed; and (c) if the approximation region is determined to be sparsely distributed, indexing the corresponding approximation region as one special node belonging to a child node of the tree data structure, together with any other sparsely distributed approximation region spaced apart by a distance less than a predetermined distance.

Abstract: An object activity modeling method which can efficiently model complex objects such as a human body is provided. The object activity modeling method includes the steps of (a) obtaining an optical flow vector from a video sequence; (b) obtaining the probability distribution of the feature vector for a plurality of video frames, using the optical flow vector; (c) modeling states, using the probability distribution of the feature vector; and (d) expressing the activity of the object in the video sequence based on state transition. According to the modeling method, in video indexing and recognition field, complex activities such as human activities can be efficiently modeled and recognized without segmenting objects.

Abstract: A method for extracting from an image a shape descriptor which describes shape features of the image is provided. The shape descriptor extracting method includes: (a) extracting a skeleton from an input image, (b) obtaining a list of straight lines by connecting pixels based on the extracted skeleton, and (c) determining the regularized list of straight lines obtained by normalizing a list of straight lines as the shape descriptor. A shape descriptor extracted according to the shape descriptor extracting method possesses information of a schematic feature of a shape included in an image. Therefore, the shape descriptor extracting method effectively extracts a local motion in the data collection of the same category, and the number of extracted shapes is not limited to the number of objects.

Abstract: An apparatus for searching a digital image and a method therefor are provided. The digital image search apparatus includes a database processing unit for segmenting an image from the database or an image which is not stored in the database into homogeneous regions and analyzing and storing the segmented regions and an image search engine unit for searching for an image having a segmented region showing a characteristic similar to the characteristic of the query region. It is possible to perform detailed and efficient searching work since it is possible to perform a region-based search by means of segmented regions and a user can perform searching by selecting the characteristic and the position of the segmented regions.

Abstract: A method is provided for indexing feature vectors within feature vector data space which includes adaptively forming approximation of the feature vectors on the basis of statistical distribution of feature vector data within the feature vector data space. The method for indexing feature vector data space of the present invention efficiently indexes vector space having high-dimensionality in which generally feature vectors are not uniformly distributed. Also, the method for indexing feature vector data space has an advantage that indexing can be easily upgraded when new feature vector data is added.

Abstract: A method for describing pattern repetitiveness which may exist within an image, is provided. The method of describing pattern repetitiveness of an image comprises: (a) projecting an image on a predetermined axis having a predetermined direction, (b) decomposing the projected image down one level, (c) increasing a threshold value until a pattern quantizing value is retained, and denoising the decomposed data, and (d) describing pattern repetitiveness of the image using the pattern quantizing value of the denoised data and the threshold value used for denoising. Because the method of describing pattern repetitiveness of an image effectively denoises the original image without damaging the pattern repetitiveness of the original image, and the pattern repetitiveness is described using the denoised data, it is possible to describe more definite pattern repetitiveness.

Abstract: A method of measuring a similarity between images is provided. The method includes the steps of (a) computing a statistical dissimilarity between the images; (b) computing a perceptual dissimilarity between the images; and (c) computing a dissimilarity between the texture features of the images based on the statistical dissimilarity and the perceptual dissimilarity. When images perceptually similar to a query image are searched and retrieved according to the above measuring method, the possibility that the retrieved images are not perceptually similar to the query image is small. Accordingly, the image searching performance can be improved when searching images having texture features similar to that of a query image.

Abstract: An indexing method of a feature vector data space, which can be used for a similarity search in a multidimensional vector space, is provided. The indexing method includes the steps of (a) determining whether at least one cell, on which feature vectors are concentrated, exists, and (b) hierarchically indexing the feature vector data space when it is determined that at least one cell, on which feature vectors are concentrated, exists in the step (a). Accordingly, the feature vector data space can be finely indexed when feature vectors are not uniformly distributed in a high-dimensional vector space.

Abstract: A texture description method in a frequency domain for extracting texture features by transforming and Gabor-filtering an input image into an image of the frequency domain, and a texture-based retrieval method thereof are provided. The texture description method in the frequency domain includes: a first step of transforming an image of a time domain into an image of the frequency domain; a second step of filtering the transformed image using a Gabor filter having N×M filter regions, where N and M are respective predetermined positive integers; a third step of extracting feature values of the image that has been Gabor-filtered in respective channels of the frequency domain division layout corresponding to the N×M filter regions of the Gabor filter; and a fourth step of deciding a texture descriptor of the image on the basis of the feature values of the image.