Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-08-27
2003-07-22
Philippe, Gims S. (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C382S107000, C382S168000
Reexamination Certificate
active
06597738
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a motion descriptor generating apparatus by using accumulated motion histogram and a method therefor, in which after motion information onto a video concept is processed in view of human perceptual characteristics, accumulated motion histogram is generated and a motion descriptor is generated by using the accumulated motion histogram, thereby being applicable to multi-media database or search system, that is, video application fields.
2. Description of the Conventional Art
Recently, due to the continuous development of expression media, delivery media, storing media and operating these media, demands for free production of large capacity of multimedia data rather than a mono media, rapid search and convenient use and reuse has a tendency toward increasing. The demands for production of free multimedia have been satisfied with according to the development of electronic expression media, and therefore, huge amount of mono-media or multimedia data become scattered on personal or public systems.
However, as the amount of multimedia data increase, time and expense required for searching data to use or reuse become proportionally increased. Therefore, in order to increase the speed and efficiency of data search, it has been researched and developed new search techniques which include the widely-known character-based search technique and have composite information attribute, thereby being suitable for efficient data search of multimedia.
In order to perform search and index of the multimedia data efficiently, it is necessary to achieve minimization of the number of information attribute which express respective media data, simplification of whole procedures and real-time processing. Furthermore, it is necessary to guarantee effectiveness and variety of information attribute to express and flexibility of search.
Further the subjective similarity and the objective similarity of search result is the important factors to evaluate the performance of search. The importance of the subjective similarity is resulted from the limitation in the representation of information characteristics which describes the media data characteristics. Therefore, even though the objective similarity is large, the validity and utility of the search are degraded if the search result as desired by users. Accordingly, it has been continuously studied to develop methods which may reflect subjective similarity to the information attribute which expresses the media data characteristics.
The difference between character-based search and multimedia data-based search, which are both widely used in the fields, is the difficulty of information attribute extraction and variety of information attribute description.
In the view of the difficulty of information attribute extraction, even though it is possible to conduct a search the character data by indexing several principal words and sentences in a document, the data size is to large and lots of media are mixed in case of multimedia so that a proper preprocessing is required in order to obtain new valid information attribute in which information attributes are organically combined each other.
The preprocessing is performed to extract valid characteristics information and to give validity to the procedure. In addition, even though a method is able to detect valid characteristics information; the method can not be put into practical in the application fields that the large amount of multimedia data should be processed in a short time or terminal systems have bad performance, if the expense for hardware (H/W) and software (S/W) in the procedure.
Now, the variety of information attribute description will be described in case of video search, as an example. The video is mixed with various media information such as images, voice, sound, etc. Therefore, after valid information attribute is extracted by preprocessing the information attribute of the respective mono-media data or the information attribute of multimedia which is mixed with more than two media, the data search in the various type be conducted by using the extracted information attribute. For example, in order to search the video, information attribute of images may be utilized and it is also possible to utilize the combined information attribute of images and voices. Therefore, the search which utilizes the various multimedia attributes is more effective than that which utilizes only a single mono-media attributes.
Nowadays, most of studies have been concentrated on the field of stop image which is easier to obtain data in multimedia data indexing and search. The still image is widely utilized in the fields of storing system such as digital electronic steel cameras and image database, transmission system such as stop image transmission device or audio graphic conference and image conference, and printing system such as a printer. The still image indexing or search is the image search method based on the content. Therefore, the importance resides on the extraction, indexing and search method of characteristics information which has a consistent characteristic with relation to the change such as rotation, scale, and translation of the color, texture and shape of the images.
The video search fields is not easy to obtain data comparing with the stop image and limited in its application due to the large capacity of data to store and process. However, owing to the rapid development of the transmission media and storing media such as discs, tapes and cd roms, the expense required for obtaining the data decreases-and owing to the tendency of minimizing of the necessary devices, the study in this fields becomes vigorous. In general, the video refers all the series of images which have sequence in the continuous time. Therefore, the video has the spatial redundancy (repetition) in an image and between images, which is the difference of the characteristics of the video images from the stop images. In the video search method, the reducdance between images may be importantly utilized in the extraction of the characteristics information.
The redundancy between video frames may be measured by using the motion degree. For example, if the redundancy is large, it means that the size of region is large and the motion between the region is small. On the other hand, if the redundancy is small, it means that the size of region is small and the motion between the region is large. At present, video compression methods, of which standardization is finished, adopt motion estimation between images (BMA-Block Matching Algorithm) for thereby improving data compression efficiency (H.261, H.263, MPEG-1, MPEG-2, MPEG-4).
In the conventional video search method, a certain temporal position (hereinafter, to be referred to “clip”) of certain units is structured on the basis of changes in the color, texture, shape and motion between images, several key frames, which represent meaning characteristics and signal characteristics of the images in the clip, are selected, and the characteristics information are extracted with relation to the information attributes of the selected key frames to perform them in indexing or search.
In the video structuring, general structure is to be a hierarchical structure which is comprised of basic unit “shot” which is a series of stop images having no temporal disconnection, “Scene” which is a series of continuous shots, having temporal and spatial continuity in the content, and “Story” which is a series of continuous scenes in the four steps of composition.
FIG. 17B
shows the general video structure. The video structuring may be achieved in the type of an event tree on the basis of signal characteristics. In the video structuring, the structuring information of the signal characteristics and the significant characteristics may exist together on the basis of correlated link.
In
FIG. 17B
, the segment tree as shown in the left side and the segment tree as shown in the right side are linked together in the direction as shown by the arro
Kim Hae-Kwang
Moon Joo-Hee
Park Cheol-Soo
Blakely & Sokoloff, Taylor & Zafman
Hyundai Curitel Inc.
Philippe Gims S.
LandOfFree
Motion descriptor generating apparatus by using accumulated... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Motion descriptor generating apparatus by using accumulated..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Motion descriptor generating apparatus by using accumulated... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3082481