Multiple feature kernel hashing for large-scale visual search

Recently hashing has become attractive in large-scale visual search, owing to its theoretical guarantee and practical success. However, most of the state-of-the-art hashing methods can only employ a single feature type to learn hashing functions. Related research on image search, clustering, and oth...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Pattern recognition Ročník 47; číslo 2; s. 748 - 757
Hlavní autori: Liu, Xianglong, He, Junfeng, Lang, Bo
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Kidlington Elsevier Ltd 01.02.2014
Elsevier
Predmet:
Analogies
Applied sciences
Compact hashing
Exact sciences and technology
Image processing
Information theory
Information, signal and communications theory
Kernels
Locality-sensitive hashing
Multiple features
Multiple kernels
Optimization
Pattern recognition
Preserves
Searching
Signal and communications theory
Signal processing
Signal representation. Spectral analysis
Signal, noise
State of the art
Telecommunications and information theory
Visual
Compact hashing
Multiple features
Locality-sensitive hashing
Multiple kernels
Performance evaluation
Automatic classification
State of the art
Similarity
Image retrieval
Information retrieval
Unsupervised classification
Signal classification
Optimization
Kernel method
Accuracy
Algorithm performance
Visual search
Hashing
ISSN:0031-3203, 1873-5142
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Recently hashing has become attractive in large-scale visual search, owing to its theoretical guarantee and practical success. However, most of the state-of-the-art hashing methods can only employ a single feature type to learn hashing functions. Related research on image search, clustering, and other domains has proved the advantages of fusing multiple features. In this paper we propose a novel multiple feature kernel hashing framework, where hashing functions are learned to preserve certain similarities with linearly combined multiple kernels corresponding to different features. The framework is not only compatible with general types of data and diverse types of similarities indicated by different visual features, but also general for both supervised and unsupervised scenarios. We present efficient alternating optimization algorithms to learn both the hashing functions and the optimal kernel combination. Experimental results on three large-scale benchmarks CIFAR-10, NUS-WIDE and a-TRECVID show that the proposed approach can achieve superior accuracy and efficiency over state-of-the-art methods. •We propose a generic multiple feature hashing framework using multiple kernels.•Visual features are implicitly mapped and concatenated to reduce complexity.•We formulate both supervised and unsupervised hashing problems in the framework.•Alternating optimization ways efficiently learn hashing functions and the kernel space.•Experiments validate the superior performances and efficiency of the proposed approach.
Bibliografia:ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 23
ISSN:0031-3203
1873-5142
DOI:10.1016/j.patcog.2013.08.022