Detailed Information on Publication Record
2015
Efficient Image Search with Neural Net Features
NOVÁK, David, Jan ČECH and Pavel ZEZULABasic information
Original name
Efficient Image Search with Neural Net Features
Authors
NOVÁK, David (203 Czech Republic, guarantor, belonging to the institution), Jan ČECH (203 Czech Republic) and Pavel ZEZULA (203 Czech Republic, belonging to the institution)
Edition
New York, Similarity Search and Applications: 8th International Conference, SISAP 2015, Glasgow, UK, October 12-14, 2015, Proceedings, p. 237-243, 7 pp. 2015
Publisher
Springer International Publishing
Other information
Language
English
Type of outcome
Stať ve sborníku
Field of Study
10201 Computer sciences, information science, bioinformatics
Country of publisher
United States of America
Confidentiality degree
není předmětem státního či obchodního tajemství
Publication form
printed version "print"
References:
Impact factor
Impact factor: 0.402 in 2005
RIV identification code
RIV/00216224:14330/15:00081692
Organization unit
Faculty of Informatics
ISBN
978-3-319-25086-1
ISSN
UT WoS
000374289600022
Keywords in English
metric indexing; deep convolutional neural network; contentbased image retrieval
Tags
International impact, Reviewed
Změněno: 28/4/2016 15:35, RNDr. Pavel Šmerk, Ph.D.
Abstract
V originále
We present an efficiency evaluation of similarity search techniques applied on visual features from deep neural networks. Our test collection consists of 20 million 4096-dimensional descriptors (320GB of data). We test approximate k-NN search using several techniques, specifically FLANN library (a popular in-memory implementation of k-d tree forest), M-Index (that uses recursive Voronoi partitioning of a metric space), and PPP-Codes, which work with memory codes of metric objects and use disk storage for candidate refinement. Our evaluation shows that as long as the data fit in main memory, the FLANN and the M-Index have practically the same ratio between precision and response time. The PPP-Codes identify candidate sets ten times smaller then the other techniques and the response times are around 500 ms for the whole 20M dataset stored on the disk. The visual search with this index is available as an online demo application. The collection of 20M descriptors is provided as a public dataset to academic community.
Links
GAP103/10/0886, research and development project |
|