Name
Abstract | ||
|---|---|---|
Recent years have witnessed a growing interesting in studying of painting images. It is obvious that there exists a deep gap between painting images and natural images, due to special characteristics of painting images. Therefore, general image classification methods are not suitable to be applied directly to the painting images. This paper demonstrates a simple, yet powerful on-line learning algorithm to classify the category of painting images. Specifically, we use the multi-features combining of local and global features as the image descriptor, and then K-means is applied to initialize the dictionary. We resort to the online learning method to optimize the dictionary which then served as a codebook for spare coding of multi-features. Finally, we facilitate the linear support vector machine to classify images. The experimental results on two painting image datasets show that, compared with the traditional image classification method, our method has improved the accuracy of image classification. What's more, from the practical viewpoint, our online learning mechanism can be also useful for many other pattern recognition tasks. Based on the research results of this paper, it can be applied in various fields such as art field and art analysis, the research of this paper provides a new way for art researchers to explore the potential of computer-aided analysis of painting works and promote the development of art research. Painting image classification could be used in social internet of things such as in museum, to make people understand the painting more in-depth. |
Year | DOI | Venue |
|---|---|---|
2017 | 10.1007/978-3-319-58697-7_29 | DISTRIBUTED, AMBIENT AND PERVASIVE INTERACTIONS, DAPI 2017 |
Keywords | Field | DocType |
Painting image classification, Multi-features, Online learning | Computer science,Painting,Coding (social sciences),Artificial intelligence,Contextual image classification,Online learning,Computer vision,Spare part,Existential quantification,Support vector machine,Algorithm,Machine learning,Codebook | Conference |
Volume | ISSN | Citations |
10291 | 0302-9743 | 2 |
PageRank | References | Authors |
0.37 | 8 | 4 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Bing Yang | 1 | 44 | 8.37 |
| Jin-liang Yao | 2 | 12 | 3.72 |
| Xin Yang | 3 | 200 | 36.16 |
| Yan Shi | 4 | 60 | 12.08 |