• Journal of Internet Computing and Services
• /
• v.10 no.6
• /
• pp.173-189
• /
• 2009

Recently, video data has attracted many interest. That is the reason why efficient indexing schemes are required to support the content-based retrieval of video data. But most indexing schemes are not suitable for indexing a high-dimensional data except Hybrid Spill-Tree. In this paper, we propose an efficient high-dimensional indexing scheme to support the content-based retrieval of video data. For this, we extend Hybrid Spill-Tree by using a newly designed clustering technique and by adopting a signature method. Finally, we show that proposed signature-based high dimensional indexing scheme achieves better retrieval performance than existing M-Tree and Hybrid Spill-Tree.

• Journal of IKEEE
• /
• v.14 no.4
• /
• pp.339-345
• /
• 2010

So far, many researches have been done to index high-dimensional feature values for fast content-based image retrieval. Still, many existing indexing schemes are suffering from performance degradation due to the curse of dimensionality problem. As an alternative, heuristic algorithms have been proposed to calculate the result with 'high probability' at the cost of accuracy. In this paper, we propose a new extendible hash-based indexing scheme for high-dimensional feature values. Our indexing scheme provides several advantages compared to the traditional high-dimensional index structures in terms of search performance and accuracy preservation. Through extensive experiments, we show that our proposed indexing scheme achieves outstanding performance.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.1
• /
• pp.70-74
• /
• 2010

As the storage capacity of the flash memories increased portable devices began to store mass amount of multimedia data on flash memory. Therefore, there has been a need for an effective data management scheme by indexing structure. Among many indexing schemes, M-tree is well known for it's suitability for multimedia data with high dimensional matrix space. Since flash memories have writing operation restriction, there is a performance limitation in indexing scheme with frequent write operation. In this paper, a new node split method with reduced write operation for m-tree indexing scheme in flash memory is proposed. According to experiments the proposed method reduced the write operation to about 7% of the original method. The proposed method will effectively construct an indexing structure for multimedia data in flash memories.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06a
• /
• pp.333-334
• /
• 2008

• Proceedings of the Korean Information Science Society Conference
• /
• 1999.10a
• /
• pp.219-221
• /
• 1999

현재 멀티미디어 응용분야에서 고차원 데이터에 대한 색인 기법이 아주 중요시 되고 있다. 특히, 인터넷의 보급으로 멀티미디어 정보에 대한 수요가 급증함에 따라 멀티미디어 객체에 대한 효율적인 색인 기술이 절실히 필요하게 되었다. 멀티미디어 객체들은 특징 벡터들로 표현이 되며, 대부분 고차원 특징 벡터를 형성하게 된다. 이러한 고차원 특징 벡터를 색인 및 검색하기 위하여 다양한 방법들이 제시되었다. 그러나, 차원이 증가할수록 검색 성능이 급격히 저하되는 dimensional curse 문제를 완전히 해결하지는 못했다. 본 논문에서는 필터링(filtering) 기법을 사용하여 개선된 고차원 색인 기법을 설계 및 구현한다.

• The KIPS Transactions:PartD
• /
• v.19D no.2
• /
• pp.151-160
• /
• 2012

Most of the distributed high-dimensional indexing structures provide a reasonable search performance especially when the dataset is uniformly distributed. However, in case when the dataset is clustered or skewed, the search performances gradually degrade as compared with the uniformly distributed dataset. We propose a method of improving the k-nearest neighbor search performance for the distributed vector approximation-tree based on the strongly clustered or skewed dataset. The basic idea is to compute volumes of the leaf nodes on the top-tree of a distributed vector approximation-tree and to assign different number of bits to them in order to assure an identification performance of vector approximation. In other words, it can be done by assigning more bits to the high-density clusters. We conducted experiments to compare the search performance with the distributed hybrid spill-tree and distributed vector approximation-tree by using the synthetic and real data sets. The experimental results show that our proposed scheme provides consistent results with significant performance improvements of the distributed vector approximation-tree for strongly clustered or skewed datasets.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.2
• /
• pp.69-77
• /
• 2009

Stream service environment demands real-time query processing for voluminous data which are ceaselessly delivered from tremendous sources. Typical R-tree based query processing technologies cannot efficiently handle such situations, which require repetitive and inefficient exploration from the tree root on every data event. However, many stream data including sensor readings show high locality, which we exploit to reduce the search space of queries to explore. In this paper, we propose a query processing scheme exploiting the locality of stream data. From the simulation, we conclude that the proposed scheme performs much better than the traditional ones in terms of scalability and exploration efficiency.

• Journal of KIISE:Databases
• /
• v.28 no.2
• /
• pp.204-216
• /
• 2001

최근 이미지 특징 벡터와 같은 고차원 벡터 데이터에 관한 색인 기법들이 많이 연구되고 있다. 하지만, 기존의 색인 기법들은 저차원의 데이터에 대해서는 검색 성능이 우수하지만, 차원이 증가함에 따라 검색 성능이 급격히 저하되는 'dimensional curse' 문제를 안고 있다. 따라서, 본 논문에서는 이러한 문제점을 최소화하기 위해 필터링을 이용한 새로운 색인 기법을 제안한다. 제안하는 셀 기반 필터링 기법은 셀 중심에서 객체까지의 거리값을 저장하여 필터링 효과를 증대시킨다. 또한 고차원 공간을 셀 단위로 분할하며, 각각의 셀을 시그니쳐로 표현한다. 검색을 수행하기 위해, 셀 기반 필터링 기법은 데이터 파일을 직접 접근하기 전에 전체 시그니쳐들을 탐색하여 필터링을 수행함으로써 후보 셀들을 얻는다. 성능 실험을 통해 제안하는 기법이 VA-파일보다 검색 시간에 있어서 약 20%의 성능 향상을 보인다.