Author:
Me Mingyu,Zhang Jiangzhou,Li Yang,Wang Shuai
Abstract
Abstract
This paper adds the spatial and semantic relations of the "visual dictionary", the longest common visual word string constraint in the image based on the BoW model. The image is represented at the vocabulary level and the phrase level at the same time. This algorithm can better represent the content of the image, thereby improving the quality of image retrieval. Experiments verify that compared with the BoW model, the accuracy of the BoW image retrieval model based on the longest common word string constraint is well optimized.
Subject
General Physics and Astronomy
Reference11 articles.
1. A comparative study on feature extraction using texture and shape for content based image retrieval;Bagri;International Journal of Advanced Science and Technology,2015
2. Video Google: A text retrieval approach to object matching in videos;Sivic,2003
3. Feature integration analysis of bag-of-features model for image retrieval;Yu;Neurocomputing,2013
4. Low-level features for retrieval based on extraction of directional binary patterns and its oriented gradients histogram;Nagaraja;Computer Science,2015
5. Image retrieval algorithm based on SIFT, K-Means and LDA;Yulei;Journal of Beijing University of Aeronautics,2014