102 Hits in 5.9 sec

A Survey on Efficient Processing of Similarity Queries over Neural Embeddings [article]

Yifan Wang
2022 arXiv   pre-print
Then we talk about recent approaches on designing the indexes and operators for highly efficient similarity query processing on top of embeddings (or more generally, high dimensional data).  ...  Finally, we investigate the specific solutions with and without using embeddings in selected application domains of similarity queries, including entity resolution and information retrieval.  ...  , in another word, using a randomly rotated orthogonal basis for the space to build a KD-tree.  ... 
arXiv:2204.07922v1 fatcat:u5osyghs6vgppnj5gpnrzhae5y

Semantic Models for the First-stage Retrieval: A Comprehensive Review [article]

Yinqiong Cai, Yixing Fan, Jiafeng Guo, Fei Sun, Ruqing Zhang, Xueqi Cheng
2021 arXiv   pre-print
Therefore, it has been a long-term desire to build semantic models for the first-stage retrieval that can achieve high recall efficiently.  ...  In this paper, we describe the current landscape of the first-stage retrieval models under a unified framework to clarify the connection between classical term-based retrieval methods, early semantic retrieval  ...  For example, the KD Tree [11] almost degenerates into brute-force search in practical scenarios.  ... 
arXiv:2103.04831v3 fatcat:6qa7hvc3jve3pcmo2mo4qsiefq

Similarity Search on Automata Processors [article]

Vincent T. Lee, Justin Kotalik, Carlo C. Del Mundo, Armin Alaghi, Luis Ceze, Mark Oskin
2017 arXiv   pre-print
At its core, similarity search is implemented using the k-nearest neighbors (kNN) algorithm, where computation consists of highly parallel distance calculations and a global top-k sort.  ...  In this paper, we present and evaluate a novel automata-based algorithm for kNN on the Micron Automata Processor (AP), which is a non-von Neumann near-data processing architecture.  ...  For simplicity, we use four hash tables for LSH and four parallel kd-trees; for kd-trees and k-means each tree traversal checks one bucket of vectors.  ... 
arXiv:1608.03175v2 fatcat:gqtchulalnea3mdasisfqz7sgq

High order pLSA for indexing tagged images

S. Nikolopoulos, S. Zafeiriou, I. Patras, I. Kompatsiaris
2013 Signal Processing  
Then, by treating images, visual features and tags as the three observable variables of an aspect model, we learn a space of latent topics that incorporates the semantics of both visual and tag information  ...  This work presents a method for the efficient indexing of tagged images. Tagged images are a common resource of social networks and occupy a large portion of the social media stream.  ...  Finally, in order to facilitate fast image matching the images were indexed using a kd-tree multidimensional indexing structure [48] that supports k-NN (nearest neighbor) queries.  ... 
doi:10.1016/j.sigpro.2012.08.004 fatcat:2bacd2atq5eivn6ths6ehftzc4

A Natural-language-based Visual Query Approach of Uncertain Human Trajectories

Zhaosong Huang, Ye Zhao, Wei Chen, Shengjie Gao, Kejie Yu, Weixia Xu, Mingjie Tang, Minfeng Zhu, Mingliang Xu
2019 IEEE Transactions on Visualization and Computer Graphics  
On the other hand, domain experts and general users prefer a natural way, such as using a natural language sentence, to access and analyze massive movement data.  ...  an effective indexing scheme.  ...  R-tree [20, 29] , KD-Tree [20, 29] and their extensions (e.g., [54, 56, 61, 79] ) are widely used to improve the performance of KNN-based queries.  ... 
doi:10.1109/tvcg.2019.2934671 pmid:31443013 fatcat:pa7womdqdrasnl6awvhoynuuhi

Content-Based Management of Human Motion Data: Survey and Challenges

Jan Sedmidubsky, Petr Elias, Petra Budikova, Pavel Zezula
2021 IEEE Access  
Digitization of human motion using skeleton representations offers exciting possibilities for a large number of applications but, at the same time, requires innovative techniques for their effective and  ...  Next, we review the techniques for evaluating similarity queries over collections of motion sequences and filtering query-relevant parts from continuous motion streams.  ...  In the pre-search step, a set of query-relevant candidate results is efficiently retrieved, e.g., using various index structures [82] , such as the binary tree [70] , kd tree [71] , or tries [77] .  ... 
doi:10.1109/access.2021.3075766 doaj:29fcaf076f214ea78579a7b373e2d130 fatcat:vdrg455hazd55jgfjn2udoadvq


F. Poux, P. Hallot, R. Neuville, R. Billen
2016 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
A review of feature detection, machine learning frameworks and database systems indexed both for mining queries and data visualisation is studied.  ...  We propose to use both point cloud properties and human knowledge through machine learning to rapidly extract pertinent information, using user-centered information (smart data) rather than raw data.  ...  Based on LoD concepts aggregating points regarding attributes, they define a new class-attached point cloud out of core rendering system by storing points in a layered multi-resolution kd-tree.  ... 
doi:10.5194/isprs-annals-iv-2-w1-119-2016 fatcat:prtjp6m6gfbs7dfvvot7lx62mu

Incremental Indexing for High-Dimensional Data using Tree Structure

R Vishnu Priya, A. Vadivel
2012 Procedia Technology - Elsevier  
A novel indexing scheme is proposed in this paper to achieve a fast response and higher precision of retrieval.  ...  For experimental purpose, the coral image database used and found that the performance of the proposed indexing scheme is encouraging.  ...  The main problem of KD-tree is that recognizing the position of the feature vector using each level of the tree.  ... 
doi:10.1016/j.protcy.2012.10.065 fatcat:ztdfvorjhfat7g2rna2lvb5eca

A scalable approach for content based image retrieval in cloud datacenter

Jianxin Liao, Di Yang, Tonghong Li, Jingyu Wang, Qi Qi, Xiaomin Zhu
2013 Information Systems Frontiers  
However, traditional search techniques only allow users to search images over exact-match keywords through a centralized index.  ...  The emergence of cloud datacenters enhances the capability of online data storage.  ...  But its data clustering and mapping are still completed in a centralized model. In Psearch, the Latent Semantic Indexing (LSI) is used to generate a semantic space.  ... 
doi:10.1007/s10796-013-9467-0 fatcat:6u5d7ps2tnfhdjo2wn4676z3pe

A review of EO image information mining [article]

Marco Quartulli, Igor G. Olaizola
2012 arXiv   pre-print
The solutions envisaged for the issues related to feature simplification and synthesis, indexing, semantic labeling are reviewed. The methodologies for query specification and execution are analyzed.  ...  ., images) in the database would imply of O(n) computations, a kd-tree approach involves a complexity O(log 2 (n)).  ...  Semantic annotation of satellite images using the Latent Dirichlet Allocation (LDA) model.  ... 
arXiv:1203.0747v2 fatcat:nwiylcsdrnhthi753xcxwxgo7e

Enriching and Clustering Short Text Using KNN

Ms. Shalika, Mr. Veepin Kumar
2021 International Research Journal on Advanced Science Hub  
So, to find out that whether two short texts are alike or not in their meaning, their binary codes need to be matched. A deep neural network is used for encoding.  ...  Additionally, we use a k-Nearest Neighbor based approach id for hashing.  ...  While using K-NN, in a very low dimensional space we can use an RP-Tree or KD-Tree to improve its performance.  ... 
doi:10.47392/irjash.2021.219 fatcat:vcbl3njzu5hdvgnrp74mptacfe

Matching Handwritten Document Images [article]

Praveen Krishnan, C.V. Jawahar
2016 arXiv   pre-print
Finally, we demonstrate the applicability of our method on a practical problem of matching handwritten assignments.  ...  We formulate the document matching problem as a structured comparison of the word distributions across two document images.  ...  In order to reduce the exhaustive matches, we use an approximate nearest neighbor search using KD trees.  ... 
arXiv:1605.05923v1 fatcat:slalf42hi5cfvdsa7oixmj2mue

Effective of Modern Techniques on Content-Based Medical Image Retrieval: A Survey

Metwally Rashad, Sameer Nooh, Ibrahem Afifi, Mohamed Abdelfatah
2022 International journal of computer science and mobile computing  
The majority of the methods already in use in CBMIR enhance the retrieval of a medical image and diseases diagnosis by reducing the issue of the semantic gap between low visual and high semantic levels  ...  This implies that a precise, efficient way of indexing and retrieving biomedical images is necessary to obtain medical images from such repositories in real-time.  ...  done by algorithms of the nearest neighbor and KD tree with the best bin first (BBF).  ... 
doi:10.47760/ijcsmc.2022.v11i03.008 fatcat:656cjypw75h43mjfflbngdouia

Scalable ranked retrieval using document images

Rajiv Jain, Douglas W. Oard, David Doermann, Bertrand Coüasnon, Eric K. Ringger
2013 Document Recognition and Retrieval XXI  
This indexing scheme is used to create the following inverted index: Each index key points to the unique ID for the document it was computed from and its associated feature vector.  ...  To minimize the storage cost and computational requirements of this matching, the SURF feature vector is reduced to 8 dimensions using PCA.  ...  The features are mapped to code words and then represented using Latent Semantic Indexing.  ... 
doi:10.1117/12.2038656 dblp:conf/drr/JainOD14 fatcat:fakph7vukvglnanvqfjadbx42a

Large-Scale Indexing, Discovery, and Ranking for the Internet of Things (IoT)

Yasmin Fathy, Payam Barnaghi, Rahim Tafazolli
2018 ACM Computing Surveys  
The primary objective of this paper is to provide a holistic overview of the state-of-the-art on indexing, discovery and ranking of IoT data.  ...  However, the existing IoT data indexing and discovery approaches are complex or centralised which hinders their scalability.  ...  Spatial data can be indexed using tree-based methods such as R-tree, B-tree and kd-tree.  ... 
doi:10.1145/3154525 fatcat:g6h4h6ezorhkda6qmgibgn4dtm
« Previous Showing results 1 — 15 out of 102 results