Yihong Gong, Hongjiang Zhang, H. C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proceedings of the International Conference on Multimedia Computing and Systems, pages 121--130. IEEE, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....hard to account even for small rotations and translations of an image. Relying on a complete segmentation implies in most domains of application a large amount of user interaction at the database population phase which is too time consuming for large image or video databases. Most systems [5, 16, 1] work with more than one index to exploit the advantages of different approaches: no spatial information, image subdivisions, automated or user assisted segmentation. The disadvantage of such an approach is that using several indices can make the combined index relatively large. Furthermore, a ....

Y. Gong, H. Zhang, H.C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proc. of the Intl. Conf. Multimedia Computing and Systems, May 1994.

....hard to account even for small rotations and translations of an image. Relying on a complete segmentation implies in most domains of application a large amount of user interaction already at the database population phase which is too time consuming for large image or video databases. Most systems [Gong et al. 1994, Zhang and Smoliar 1994, Ashley et al. 1995] work with more than one index to exploit the advantages of different approaches: no spatial information, image subdivisions, automated or user assisted segmentation. The disadvantage of such an approach is that using several indices can make the ....

Y. Gong, H. Zhang, H.C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proc. of the Intl. Conf. Multimedia Computing and Systems, May 1994.

....multimedia databases increasingly common. To allow the end user accessing the enormous amount of information contained in an image database, existing database management methods have to be extended. Image contents representation by indexing is a promising way to cope with image retrieval problems [6] [4] Most of the current systems work in the same way: From each image region a feature vector is extracted. The set of all these vectors is organized in a database. The systems differ mainly in the feature that they extract, the organization of the access structure and the comparison of the ....

Y. Gong, H. Zhang, H.C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proc. of the Intl. Conf. Multimedia Computing and Systems, May 1994.

....for accessing multimedia material, such as video and still images, using their media speci c features. In particular, several research papers and tools have been presented for image retrieval based on low level visual features, such as color and luminance, as image descriptors ( 3] 5] 6] 7] [8], 9] 10] 15] 16] 17] 19] 22] This paper considers how an ecient and e ective system for image retrieval can be based on the statistics of such low level features. A novel de nition of histogram capacity curve taking into account the density distribution of histograms in the ....

Y. Gong, H. Zhang, H. C. Chuan, and M. Sakauchi. An Image Database System with Content Capturing and Fast Image Indexing Abilities. In Proceedings of IEEE International Conference on Multimedia Computing and Systems, pages 121-130, Boston, May 1994.

....Computer Vision community is the design and development of efficient tools for accessing multimedia material using their media specific features. In particular, several research papers and tools have been presented for image retrieval based on low level visual features such as color and luminance ([2, 3, 5, 6, 7, 8, 10]) This paper considers how an efficient and effective system for image retrieval can be based on the use of histograms of different features as image descriptors. Several issues are considered: ffl how descriptor effectiveness should be assessed, ffl how histograms should be compared, ffl how ....

Y. Gong, H. Zhang, H. C. Chuan, and M. Sakauchi. An Image Database System with Content Capturing and Fast Image Indexing Abilities. In Proceedings of IEEE International Conference on Multimedia Computing and Systems, pages 121--130, Boston, May 1994.

....modify his or her query accordingly. The distinct feature of CAETIIML (http: www.videolib.princeton.edu test retrieve) built at Princeton University, is its combination of the on line similarity searching and o line subject searching [169] More Image Retrieval systems can be found in [48, 9, 130, 153, 50, 167, 32, 152, 100]. 5. FUTURE RESEARCH DIRECTIONS From the above review, we can see that many advances have been made in various research aspects, including visual feature extraction, multi dimensional indexing, system design, etc. 147, 5, 65, 64, 66, 68, 69] However, there are still many open research issues ....

Yihong Gong, HongJiang Zhang, H.C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proc. IEEE Int. Conf. on Image Proc., 1994.

....interrupt, will identify the feature (called F ) in state s i that the user is referring to. This may require signal processing and or statistical pattern recognition techniques which are beyond the scope of this paper, but which are studied in other papers (e.g. Niblack et al. 23] Gong et al. [11], and Gupta et al. 14] The situation we are dealing with is one where the current media event is (s 1 ; s n ) and the user wishes to obtain further details on one or more features of the current media event. Query I. Suppose the user wishes to find all states (irrespective of the ....

Y. Gong, H. Zhang, H.C. Chuan and M. Sakauchi. (1994) An Image Database System with Content Capturing and Fast Image Indexing Abilities, Proc. 1994 Intl. Conf. on Multimedia Computing and Systems, pps 121-130, IEEE Press.

....discuss some other related work in the areas of content based image retrieval, image subregion querying, object localization, and cut detection. 2.1. Content based Image Retrieval Several recently proposed schemes incorporate spatial information about colors to improve upon the histogram method [18, 40, 41, 35, 11, 32, 31]. One common approach is to divide images into 4 Fig. 1. Two similar images with different histograms. subregions and impose positional constraints on the image comparison. Another approach is to augment the histogram with some spatial information. Hsu et al. 18] select two representative ....

Yihong Gong et al. "An image database system with content capturing and fast image indexing abilities," Intl. Conf. on Multimedia Comp & Systems, pp. 121-- 130, 1994.

....recorded voices. The aforementioned applications of queries by example in audio databases make it amply clear that we need novel and efficient ways for indexing and similarity searching in multimedia databases. Although there have been several efforts toward the development of databases of images [2, 3, 4, 5, 6, 7, 9, 10] and video [11, 12, 13] not much attention has been given to audio data [16, 14, 15] Audio data could serve as an independent data type or as part of multimedia data. A transform based indexing scheme and two sequential search algorithms for audio databases has been proposed in another paper ....

Gong, Y., et.al. `An Image Database System with Content Capturing and Fast Image Indexing Abilities ', Proc. Int'l Conf. on Multimedia Computing and Systems, 1994, pp121-130.

....supports the unformatted view of the pictures. Therefore, there is a considerable scope for posing ad hoc queries on the picture database and to facilitate content based picture retrieval. However, it is well recognized that an automated approach to object recognition is a very dicult problem [53, 111, 156, 59]. Furthermore, the problem is worsened if we assume that the objects to be recognized originate from disparate domains. Approaches that require the query processor to operate on the physical representation of pictures to facilitate content based retrieval further render themselves unsuitable for ....

....one hand, the amount of human eort involved is minimal, and on the other hand, the objects are recognized from disparate domains and features are automatically calculated from noisy images. Various issues related to (iconic) indexing of images and associated retrieval strategies can be found in [36, 34, 35, 40, 18, 17, 108, 109, 85, 9, 21, 33, 37, 51, 91, 90, 59]. 4.5 Attribute Similarity Most of the data types or value sets associated with picture attributes are likely to be dierent from those of the conventional data types such as Integer, Real, and String. Therefore, similarity measures for attributes based on the Euclidean distance are clearly ....

Y. Gong et al. An image database system with content capturing and fast image indexing abilities. In International Conference on Multimedia Systems and Computing, pages 121-139, Boston, MA., 1994.

....data involves range queries which retrieve all the data objects intersecting with the user defined hyper rectangle(search rectangle) 4] Visual data such as image is one of the multimedia data. Features acquisition of an image includes image segmentation, region detection and image histogram[3]. Features such as color, texture, shape[6] and so on are used for indexing. A feature vector is typically extracted for each image. Due to the high complexity of this representation for multimedia data, the DBMS should be able to handle this large index dimensions. Since the spatial data in ....

Yihong Gong, Hongjiang Zhang, H.C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilites. In Int. Conf. on Multimedia Computing and Systems, pages 121--30, 1994.

....be considered as a special case of media instances defined in their work. The above work provides a general framework for multimedia systems but does not give any complete deductive system. More recent works, such as the QBIC (Query By Image Content) NBE93, Ni93] or Photobook [PPS94] and others [GZCS94], consider similarity based retrieval of pictures using indices. However, their retrieval is primarily based on color variations, shapes of objects and texture. They do not consider spatial relationships. 3 Notation and Definitions Objects and Pictures We assume that each object has a unique ....

Gong Y. et al, An Image Database System with Content Capturing and Fast Image Indexing Abilities IEEE Multimedia Conference, 1994.

....modify his or her query accordingly. The distinct feature of CAETIIML (http: www.videolib.princeton.edu test retrieve) built at Princeton University, is its combination of the on line similarity searching and off line subject searching [60] More Image Retrieval systems can be found in [49, 10, 131, 154, 51, 168, 33, 153, 101]. 5. FUTURE RESEARCH DIRECTIONS From the above review, we can see that many advances have been made in various research aspects, including visual feature extraction, multi dimensional indexing, system design, etc. 148, 6, 66, 65, 67, 69, 70] However, there are still many open research issues ....

Yihong Gong, HongJiang Zhang, H.C. Chuan, and M.Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proc IEEE 1994, 1994.

....retrieve false positives, i.e. images with completely different content which just happen to have a similar color composition as the query image. Thus, in practice it is necessary to combine color indexing with texture and or shape indexing methods (see [ Niblack et al. 1993, Pentland et al. 1994, Gong et al. 1994 ] Even if texture and shape indexing methods improve, we believe that color indexing will retain its importance as a computationally simple and fast filter whose output is then processed by computationally more intensive methods. In this context the major challenge of new color indexing methods ....

Y. Gong, H. Zhang, et al. An image database system with content capturing and fast image indexing abilities. In Proc. of the IEEE International Conference on Multimedia Computing and Systems, pages 121--130, May 1994. Boston, Mass.

....in the management of large amount of image data pose new challenges for image segmentation. The emerging importance of image databases creates the need for new methods to access and to search data. Accordingly, there are strong ongoing research efforts on content based image retrieval [7] [9] [19] 16] In some of these systems, like in [16] a partition of the image into meaningful regions is necessary. This calls for unsupervised segmentation or region labeling algorithms, where neither training of the classifier nor human interaction takes place. Because database systems deal ....

Y. Gong, H. Zhang, H. Chuan, and M. Sakauchi. An image database system with content capturing and fast image

....virtually impossible to reconstruct the original image from its signature. Previous approaches to content based image querying have applied such properties as color histograms [Swa93] texture analysis [KZL94] and shape features like circularity and major axis orientation of regions in the image [GZCS94] as well as combinations of these techniques. One of the most notable systems for querying by image content, called QBIC, was developed at IBM [NBE 93] and is now available commercially. The emphasis in QBIC is in allowing a user to compose a query based on a variety of different visual ....

Yihong Gong, Hongjiang Zhang, H. C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proceedings of the International Conference on Multimedia Computing and Systems, pages 121--130. IEEE, 1994.

....for that image might be: red: 4 blue: 2 green: 3 Color histograms are invariant to slight modifications of position and scaling and hence provide a more accurate measure than pixel to pixel matching. Color histograms have been used in several prototype systems for indexing and retrieval of images [5, 8, 10]. They have also been used for automatic segmentation of video data into distinctive scenes to enable retrieval later on [2] The technique we have proposed takes as input an image and its loss induced version and outputs the col diff (color difference) between them. The color difference is ....

Y. Gong, H. Zhang, H. C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proceedings of the International Conference on Multimedia Computing and Systems; Boston, MA, USA, pages 121--130, Boston, MA, USA, May 1994. IEEE Computer Society Press.

....of content based image retrieval (CBIR) systems is mainly limited by the gap between low level features and high level semantic concepts. In order to reduce this gap, two approaches have been widely used: regionbased features to represent the focus of the user s perceptions of image content [4, 8, 9, 14, 19, 20, 23, 29, 36, 42] and relevance feedback (RF) to learn the user s intentions [2, 13, 18, 21, 28, 32, 34, 39, 41] Contrasting to traditional approaches [12, 24, 25, 31] which compute global features of images, the region based methods extract features of the segmented regions and perform similarity comparisons ....

....issues should be addressed: 1. How to compare two images, i.e. the definition of the image similarity measure. 2. How to make it scalable 3. How to make it interactive, i.e. the strategy of relevance feedback. For the first issue, a straightforward solution adopted by most early systems [4, 8, 19, 20, 29, 39] is to use individual region toregion similarity. When using such systems, the users are forced to select a limited number of regions from the query image in order to start a query session. As discussed in [36] due to the uncontrolled nature of the images available, automatically and precisely ....

[Article contains additional citation context not shown here]

Gong, Y., Zhang, H.J., Chuan, H. C., and Sakauchi, M., "An Image Database System with Content Capturing and Fast Image Indexing Abilities". In Proceedings of IEEE International Conference on Multimedia Computing and Systems, pages 121-130, Boston, May 1994.

....In addition, qualitative research by Rodden (1999) suggests that users are likely to find searching for photos based on the text of their annotations as a more useful and likely route in future, computer aided image databases. Currently, most of the image database systems employ manual annotation (Gong et al. 1994). That is, users enter some descriptive keywords when the images are loaded registered browsed. Although manual annotation of image content is considered a best case in terms of accuracy, since keywords are selected based on human determination of the semantic content of images, it is a labor ....

Gong Y, Zhang H, Chuan HC and Sakauchi M (1994) An image database system with content capturing and fast image indexing abilities. In: Proceedings of IEEE Int. Conf. on Multimedia Computing and Systems, 1994.

No context found.

Yihong Gong, Hongjiang Zhang, H. C. Chuan, and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. In Proceedings of the International Conference on Multimedia Computing and Systems, pages 121--130. IEEE, 1994.

No context found.

Y. Gong, H. Zhang, H.C. Chuan, and M. Sakauchi. An image database system with content capturingand fast image indexingabilities. In Proceedings of International Conference on Multimedia Computing and Systems, Boston, May 1993.

No context found.

Y. Gong, H. Zhang, H.C. Chuan and M. Sakauchi. An image database system with content capturing and fast image indexing abilities. Intl. Conf. on Multimedia Comp &Systems, pp. 121--130, 1994.

No context found.

Y. Gong, H.J. Zhang, H.C. Chuan, and M. Sakauchi, "An Image Database System with Content Capturing and Fast Image Indexing Abilities," Proceedings of the IEEE Int'l. Conf. on Multimedia Computing and Systems, pp. 121-130, 1994.

No context found.

Yihong Gong, et. al.: An Image Database System with Content Capturing and Fast Image Indexing Abilities, 0-8186-55-5#94, 1994, IEEE.

No context found.

Yihong Gong, et. al.: An Image Database System with Content Capturing and Fast Image Indexing Abilities, 0-8186-55-5/94, 1994, IEEE.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC