Title :
Autoencoder: Approach to the reduction of the dimension of the vector space with controlled loss of information
Author :
Akinin, Maxim V. ; Akinina, Natalya V. ; Taganov, Alexandr I. ; Nikiforov, Michael B.
Author_Institution :
Dept. of Space Technol., RSREU, Ryazan, Russia
Abstract :
The different ways of describing the characteristics of the image. The application of autoencoder for image classification. The results of experiments showing the effectiveness of autoencoder for solving pattern classification.
Keywords :
image classification; image coding; vectors; autoencoder; dimension reduction; image classification; loss of information; pattern classification; vector space; Accuracy; Biological neural networks; Brightness; Computer architecture; Pattern recognition; Principal component analysis; Haralik´s textural features; Laws energy characteristics; autoencoder; classification; image texture; pattern recognition; principal component analysis;
Conference_Titel :
Embedded Computing (MECO), 2015 4th Mediterranean Conference on
Conference_Location :
Budva
Print_ISBN :
978-1-4799-8999-7
DOI :
10.1109/MECO.2015.7181894
