Open AccessIntegrated SURF and Spatial Augmented Color Feature Based Bovw Model with Svm for Image Classification
Author(s) -
M. Soniya,
Pallikonda Sarah Suhasini
Publication year - 2019
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.f7900.088619
Subject(s) - bag of words model in computer vision , artificial intelligence , pattern recognition (psychology) , feature (linguistics) , computer science , color histogram , visual word , feature vector , histogram , bag of words model , cluster analysis , computer vision , support vector machine , histogram of oriented gradients , color image , image (mathematics) , image retrieval , image processing , linguistics , philosophy
In this paper, Bag-of-visual-words (BoVW) model with Speed up robust features (SURF) and spatial augmented color features for image classification is proposed. In BOVW model image is designated as vector of features occurrence count. This model ignores spatial information amongst patches, and SURF Feature descriptor is relevant to gray images only. As spatial layout of the extracted feature is important and color is a vital feature for image recognition, in this paper local color layout feature is augmented with SURF feature. Feature space is quantized using K-means clustering for feature reduction in constructing visual vocabulary. Histogram of visual word occurrence is then obtained which is applied to multiclass SVM classifier. Experimental results show that accuracy is improved with the proposed method.