## Multi-Linearization Data Structure for Image Browsing (1999)

Venue: | In SPIE { The International Society for Optical Engineering |

Citations: | 11 - 2 self |

### BibTeX

@INPROCEEDINGS{Craver99multi-linearizationdata,

author = {Scott Craver and Boon-lock Yeo and Minerva Yeung},

title = {Multi-Linearization Data Structure for Image Browsing},

booktitle = {In SPIE { The International Society for Optical Engineering},

year = {1999},

pages = {155--166}

}

### OpenURL

### Abstract

Image search has been actively studied in recent years. On the other hands, image browsing has received little attention. Image browsing refers to the process of presenting some forms of overview or summary of the image relationships, thus facilitating a user to navigate across the data set and find images of interests. In this paper, we present a new data structure built on the multi-linearization of image attributes for efficient organization of the data set and fast visual browsing of the images. We describe new techniques for multi-linearization based on multiple space-filling curves and hierarchical clustering techniques. In addition to providing fast navigation, our proposed data structure allows computationally efficient insertion and deletion of images from the data set. We then present a novel image navigator and browser built on dual-linearization data structure and intuitive presentation of image relevance and relationships, demonstrate the image navigation process, and repo...

### Citations

2147 |
Dubes: "Algorithms for Clustering data
- Jain, Richard
- 2005
(Show Context)
Citation Context .... Clustering algorithms Another method to achieve linearization is through the use of clustering and the traversal of a cluster tree structure. Bottomup clustering technique such as the Complete-Link =-=[13]-=- creates a cluster tree structure by first grouping closest neighbors, then repeatedly joining closest clusters until one cluster remains. On way to achieve linearization of the data set can then be a... |

824 |
Generalized fisheye views, in
- Furnas
- 1986
(Show Context)
Citation Context ...e direction to imply the underlying image relationships. Such adaptation has precedence. The fish-eye view was proposed for presenting files in a file system to indicate their relative sizes, etc. in =-=[14]-=-. A compact and intuitive presentation for video browsing was proposed in [15] in which a video poster was constructed by mapping representation frames, based on the relative dominance values computed... |

272 |
Space-Filling Curves
- Sagan
- 1994
(Show Context)
Citation Context ... 2.1. Space-filling curves A curve is the continuous image of a closed real interval. A space-filling curve is a curve which possesses area, volume, hypervolume, or some other positive Jordan content =-=[6]-=-. One popular example is the Hilbert Curve, which can be approximated through self-similar polygons. Figure 1 shows several examples in two and three dimensions. In the figure, the 1st, 2nd, 3rd and 4... |

140 | Fractals for secondary key retrieval - Faloutsos, Roseman - 1989 |

104 |
Video visualization for compact presentation and fast browsing of pictorial content
- Yeung, Yeo
- 1997
(Show Context)
Citation Context ...recedence. The fish-eye view was proposed for presenting files in a file system to indicate their relative sizes, etc. in [14]. A compact and intuitive presentation for video browsing was proposed in =-=[15]-=- in which a video poster was constructed by mapping representation frames, based on the relative dominance values computed from analysis, into a graphic layout of variable grid sizes to present the un... |

74 | Video Abstracting
- Lienhart, Pfeiffer, et al.
- 1997
(Show Context)
Citation Context ...ns of interest. Browsing has been better explored and understood in video research, studied under the topics “video browsing”, “video visualization”, “video abstracting” or “video summar=-=ization” (see [1,2]-=- and references there). Video browsing, in general, can be divided into three steps: (i) analyze the video to obtain structural relationships among the various units (e.g., shots and scenes), (ii) det... |

65 | The earth mover’s distance, multi-dimensional scaling, and color-based image retrieval
- Rubner, Guibas, et al.
- 1997
(Show Context)
Citation Context ...means that the image searching paradigm may not be suitable for navigation because one can easily be trapped in some clusters of images. There are a few earlier works on image browsing. Rubner et al. =-=[3]-=- applied a form of dimensionality reduction called Multidimensional Scaling (MDS) to map images to two dimensional planes. Users can then examine the collections of images mapped onto the planes for a... |

38 |
Retrieving and Visualizing Video
- Yeo, Yeung
- 1997
(Show Context)
Citation Context ...ns of interest. Browsing has been better explored and understood in video research, studied under the topics “video browsing”, “video visualization”, “video abstracting” or “video summar=-=ization” (see [1,2]-=- and references there). Video browsing, in general, can be divided into three steps: (i) analyze the video to obtain structural relationships among the various units (e.g., shots and scenes), (ii) det... |

29 | Similarity pyramids for browsing and organization of large image databases
- Chen, Bouman, et al.
- 1998
(Show Context)
Citation Context ...ap images to two dimensional planes. Users can then examine the collections of images mapped onto the planes for a quick view of the distribution of the images. The image browsing work by Chen et al. =-=[4]-=- organizes an image collection into a hierarchical quad-tree. The organization is based on bottom-up clustering of the images. The quad-tree structure is then used for browsing of the image data at di... |

17 | Space filling curves and their use in the design of geometric data structures - Asano, Ranjan, et al. - 1997 |

12 |
Convergence with Hilbert’s space filling curve
- Butz
- 1969
(Show Context)
Citation Context ...lar is based on repeated division of the square / cube / hypercube / etc. into halves, a procedure amenable to binary calculation; and can be efficiently computed in an arbitrary number of dimensions =-=[11,12]-=-. For our purposes, a space-filling curve allows us to order points in a discrete space, in such a way that nearby points in the ordering are also nearby in the original data space. This means that a ... |

9 | Space-Filling Curves and a Measure of Coherence - Voorhies - 1991 |

4 | An interactive analysis method for multi-dimensional images using a hilbert curve - Kamata, Kawaguchi, et al. - 1995 |

2 |
A Peano Curve Generation Algorithm
- Musgrave
- 1991
(Show Context)
Citation Context ...lar is based on repeated division of the square / cube / hypercube / etc. into halves, a procedure amenable to binary calculation; and can be efficiently computed in an arbitrary number of dimensions =-=[11,12]-=-. For our purposes, a space-filling curve allows us to order points in a discrete space, in such a way that nearby points in the ordering are also nearby in the original data space. This means that a ... |

1 |
Efficiency through taking up space: Spacing-filling curves and their use
- Craver, Yeo
- 1997
(Show Context)
Citation Context ... two- or ten- or six-hundred-dimensional Euclidean space depending only on the orderings we pick. A preliminary study of this problem in two and three dimensional space was reported by Craver and Yeo =-=[5]-=-. In practice, we have a database structure with many advantages, including: Ease of construction and maintenance Inserting a new item into the database could be as simple as determining its position ... |