• The KIPS Transactions:PartD
• /
• v.13D no.4 s.107
• /
• pp.455-462
• /
• 2006

Recently, the filtering approach using vector approximation such as VA-file[1] or LPC-file[2] have been proposed to support similarity search in high dimensional data space. This approach filters out many irrelevant vectors by calculating the approximate distance from a query vector using the compact approximations of vectors in database. Accordingly, the total elapsed time for similarity search is reduced because the disk I/O time is eliminated by reading the compact approximations instead of original vectors. However, the search time of the VA-file or LPC-file is not much lessened compared to the brute-force search because it requires a lot of computations for calculating the approximate distance. This paper proposes a new bitmap index structure in order to minimize the calculating time. To improve the calculating speed, a specific value of an object is saved in a bit pattern that shows a spatial position of the feature vector on a data space, and the calculation for a distance between objects is performed by the XOR bit calculation that is much faster than the real vector calculation. According to the experiment, the method that this paper suggests has shortened the total searching time to the extent of about one fourth of the sequential searching time, and to the utmost two times of the existing methods by shortening the great deal of calculating time, although this method has a longer data reading time compared to the existing vector approximation based approach. Consequently, it can be confirmed that we can improve even more the searching performance by shortening the calculating time for filtering of the existing vector approximation methods when the database speed is fast enough.

• Journal of Information Technology and Architecture
• /
• v.10 no.1
• /
• pp.79-91
• /
• 2013

This paper proposes a systematic methodology that could be used to decide which one shows the best performance among space filling curves (SFCs) in applying lower-dimensional transformations to histogram sequences. A histogram sequence represents a time-series converted from an image by the given SFC. Due to the high-dimensionality nature, histogram sequences are very difficult to be stored and searched in their original form. To solve this problem, we generally use lower-dimensional transformations, which produce lower bounds among high dimensional sequences, but the tightness of those lower-bounds is highly affected by the types of SFC. In this paper, we attack a challenging problem of evaluating which SFC shows the better performance when we apply the lower-dimensional transformation to histogram sequences. For this, we first present a concept of spatial locality, which comes from an intuition of "if the entries are adjacent in a histogram sequence, their corresponding cells should also be adjacent in its original image." We also propose spatial locality preservation metric (slpm in short) that quantitatively evaluates spatial locality and present its formal computation method. We then evaluate five SFCs from the perspective of slpm and verify that this evaluation result concurs with the performance evaluation of lower-dimensional transformations in real image matching. Finally, we perform k-NN (k-nearest neighbors) search based on lower-dimensional transformations and validate accuracy of the proposed slpm by providing that the Hilbert-order with the highest slpm also shows the best performance in k-NN search.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2001.05a
• /
• pp.265-268
• /
• 2001

In this paper, we designed one-dimensional VLSI array with high speed processing in Fractal image compression. fractal image compression algorithm partitions the original image into domain blocks and range blocks then compresses data using the self similarity of blocks. The image is partitioned into domain block with 50% overlapping. Domain block is reduced by averaging the original image to size of range block. VLSI array is trying to search the best matching between a range block and a large amount of domain blocks. Adjacent domain blocks are overlapped, so we can improve of each block's processing speed using the reuse of the overlapped data. In our experiment, proposed VLSI array has about 25% speed up by adding the least register, MUX, and DEMUX to the PE.

• Journal of KIISE
• /
• v.41 no.11
• /
• pp.900-910
• /
• 2014

Multi-view video is obtained by capturing one three-dimensional scene with many cameras at different positions. The computational complexity of multi view video coding increases in proportion to the number of cameras. To reduce computational complexity and maintain the image quality, improved prediction structure and motion estimation method is proposed in this paper. The proposed prediction structure exploits an average distance between the current picture and its reference pictures. The proposed prediction structure divides every GOP into several groups to decide the maximum index of hierarchical B layer and the number of pictures of each B layer. And the proposed motion estimation method uses a hierarchical search strategy. This strategy method consists of modified diamond search pattern, progressive diamond search pattern and modified raster search pattern. Experiment results show that the complexity reduction of the proposed prediction structure and motion estimation method over JMVC (Joint Multiview Video Coding) reference model using hierarchical B pictures of Fraunhofer-HHI and TZ search method can be up to 40~70% while maintaining similar video quality and bit rates.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.4B
• /
• pp.245-257
• /
• 2012

As one of the most challenging tasks in designing the Internet routers, packet classification is required to achieve the wire-speed processing for every incoming packet. Packet classification algorithm which applies binary search on trie levels to the area-based quad-trie is an efficient algorithm. However, it has a problem of unnecessary access to a hash table, even when there is no node in the corresponding level of the trie. In order to avoid the unnecessary off-chip memory access, we proposed an algorithm using Bloom filters along with the binary search on levels to multiple disjoint tries. For ACL, FW, IPC sets with about 1000, 5000, and 10000 rules, performance evaluation result shows that the search performance is improved by 21 to 33 percent by adding Bloom filters.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.8
• /
• pp.2298-2309
• /
• 2000

This paper proposes the multidimensioa! nested attribute indexing techniques (MD- NAI) in object-oriented databases using a multidimensional index structure. Since most conventional indexing techniques for object oriented databases use a one-dimensional index stnlcture such as the B-tree, they do not often handle complex qUlTies involving both nested attributes and class hierarchies. We extend a tunable two dimensional class hierachy indexing technique(2D-CHI) for nested attributes. The 2D-CHI is an indexing scheme that deals with the problem of clustering ohjects in a two dimensional domain space that consists of a kev attribute dOI11'lin and a class idmtifier domain for a simple attribute in a class hierachy. In our extended scheme, we construct indexes using multidimensional file organizations that include one class identifier domain per class hierarchy on a path expression that defines the indexed nested attribute. This scheme efficiently suppoI1s queries that involve search conditions on the nested attribute represcnted by an extcnded path expression. An extended path expression is a one in which a class hierarchy can be substituted by an indivisual class or a subclass hierarchy in the class hierarchy.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.12
• /
• pp.3469-3476
• /
• 1999

In spatial database based on spatial data structures, instead of one-dimensional indexing structure, new indexing structure which corresponds to multi-dimensional features of spatial objects is required. In order to meet those requirements, in this paper we proposed new indexing structure for efficient retrieval of spatial database by carrying through the feature analysis of conventional multi-dimensional indexing structures. To improve the sequential search method of supernodes in the conventional X-tree and to reduce the retrieval time in case of generating the huge supernode, we proposed a indexing structure integrating the kd-tree based on point index structure into the X-tree. We implemented the proposed indexing structure and analyzed its retrieval time according to the dimension and distribution of experimental data.

• International Journal of Computer Science & Network Security
• /
• v.24 no.5
• /
• pp.79-88
• /
• 2024

The information retrieval domain deals with the retrieval of unstructured data such as text documents. Searching documents is a main component of the modern information retrieval system. Locality Sensitive Hashing (LSH) is one of the most popular methods used in searching for documents in a high-dimensional space. The main benefit of LSH is its theoretical guarantee of query accuracy in a multi-dimensional space. More enhancement can be achieved to LSH by adding a bit to its steps. In this paper, a new Dynamic Locality Sensitive Hashing (DLSH) algorithm is proposed as an improved version of the LSH algorithm, which relies on employing the hierarchal selection of LSH parameters (number of bands, number of shingles, and number of permutation lists) based on the similarity achieved by the algorithm to optimize searching accuracy and increasing its score. Using several tampered file structures, the technique was applied, and the performance is evaluated. In some circumstances, the accuracy of matching with DLSH exceeds 95% with the optimal parameter value selected for the number of bands, the number of shingles, and the number of permutations lists of the DLSH algorithm. The result makes DLSH algorithm suitable to be applied in many critical applications that depend on accurate searching such as forensics technology.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.6
• /
• pp.21-31
• /
• 2015

The Internet users want to get real-time services for various multi-media applications. Network traffic rate has been rapidly increased, and data amounts that the Internet has to carry have been exponentially increased. A packet is the basic unit in transferring data at the Internet, and packet classification is one of the most challenging functionalities that routers should perform at wire-speed. Among various known packet classification algorithms, area-based quad-trie (AQT) algorithm is one of the efficient algorithms which can lookup five header fields simultaneously. As a representative space decomposition algorithm, the AQT requires a small amount of memory in storing classification rules, but it does not provide high-speed classification performance. In this paper, we propose a new packet classification algorithm by applying a binary search for the codewords of the AQT to overcome the issue of the AQT. Throughout simulation, it is shown that the proposed algorithm provides a better performance than the AQT in the number of rule comparisons with each input packet.

• Proceedings of the KSME Conference
• /
• 2000.11b
• /
• pp.304-308
• /
• 2000

This paper presents a numerical optimization method to design geometric shape of streamwise periodic ribs mounted on one of the principal walls to enhance turbulent heat transfer in a rectangular channel flow. The golden section method is used for the one dimensional search. The optimization is based on Wavier-Stokes analysis of turbulent forced convection with $k-{\varepsilon}$ turbulence model. The width-to-height ratio of a rib is chosen as a design variable. The object function is defined as an inverse of average Nusselt number. An optimum shape of the rib has been obtained with reasonable computing time.