Search by:
Year of publication
Author name
Paper title
Formation of Feature Spaces and their Extraction from the Images for Documents Indexing
Full text (PDF)
UDC: 004.8:004.93
Publication Language: Russian
Stuc. intelekt. 2012; 17(3):151-160
Abstract: The original algorithms of image analysis for formation of feature vector are proposed. Feature vector is formed from the both type of features, geometric features and secondary complex features compiled on the basis of the logical rules. Feature vector is used to find areas with specific information on the graphic images of documents. Simulation on a test set of images of documents has shown the stability and efficiency of the algorithms.
Keywords: indexing, feature extraction, machine learning.
References:
View full text (PDF)