Title :
ART based clustering of bag-of-features for image classification
Author :
Roy, Kaushik ; Rao, G.S.V.R.K.
Author_Institution :
Cognizant Technol. Solutions, Chennai, India
Abstract :
This paper proposes Adaptive Resonance Theory (ART) based clustering of bag-of-features for image classification. ART is an unsupervised clustering technique. Bag-of-features along with SURF signatures is one of the widely used techniques used in CBIR systems and in other computer vision tasks like classification, recognition etc. KMeans clustering has been one of the traditional techniques used for clustering. Unlike KMeans the number of clusters is not required to be pre-specified in ART based clustering. This advantage of ART has been exploited in this paper for clustering of bag-of-features.
Keywords :
content-based retrieval; feature extraction; image classification; image retrieval; pattern clustering; unsupervised learning; ART based clustering; CBIR systems; KMeans clustering; adaptive resonance theory; bag-of-features; computer vision; image classification; unsupervised clustering technique; Feature extraction; Subspace constraints; Support vector machine classification; Training; Visualization; Vocabulary; ART; SURF; SVM; accuracy; bag-of-features; classification; clustering; precision; recall;
Conference_Titel :
Image and Signal Processing (CISP), 2012 5th International Congress on
Conference_Location :
Chongqing
Print_ISBN :
978-1-4673-0965-3
DOI :
10.1109/CISP.2012.6470016
