To search, Click below search items.

 

All Published Papers Search Service

Title

Machine Learning based Oil Painting Authentication and Features Extraction

Author

Israa Abdullah Albadarneh, Ashraf Ahmad

Citation

Vol. 17  No. 1  pp. 8-17

Abstract

After the amount of the available artwork in digital form is being increased, the need of finding simple and costless way to authenticate paintings is being more important. The process of artworks authentication is done using different analysis ways manually by experts or automatically by computer processing. In this paper, oil paintings authentication system using digital image processing techniques and algorithms was proposed. Features were extracted from color and texture. Machine learning methods were used to classify the tested painting on original or forgery, based on rules from the mentioned extracted features. Five different tests and two datasets were used to evaluate the proposed system. The first dataset was used on previous work, and the second was built on this research. Color and texture features were extracted from both datasets. Two classifiers were used to study the effect of classification method on the accuracy of the authentication results. Results show an improvement on the classification accuracy using the proposed system compared with previous works.

Keywords

Painting authentication, Feature extraction, Machine Learning

URL

http://paper.ijcsns.org/07_book/201701/20170102.pdf