Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Algorithms fractal image compression based on self-organizing artificial neural network

Keywords:

K.I. Sviridenkov – Ph.D. (Eng.), Associate Professor, Department of Computer Engineering, Branch of National Research University «Moscow Power Engineering Institute» in Smolensk. E-mail: sv-k-i@mail.ru O.V. Gorchakova – Master of Science, Department of Computer Engineering, Branch of National Research University «Moscow Power Engineering Institute» in Smolensk. E-mail: ole4ka_140492@mail.ru


A algorithm covering ranking blocks domain based on self-organizing artificial neural networks. Main idea of the algorithm: select the next rank block, a feature vector of this block is input to the self-organizing artificial neural network, the output of the network obtain a list of domain blocks, bust which is produced as a standard algorithm for fractal image compression. Unlike existing algorithms, this algorithm reduces the number of searches of domain blocks, thereby reducing the running time. An algorithm for fractal image compression based on self-organizing artificial neural network image. For ease of processing domain and rank blocks of the neural network with each block is assigned a characteristic vector. As a characteristic vector selected vector, consisting of two components: the standard deviation and inter-pixel contrast. Unlike existing algorithms, this algorithm has a higher rate of image compression due to the introduction of a self-organizing algorithm of artificial neural network. For compression algorithm need trained self-organizing artificial neural network as a network in this paper uses a Kohonen self-organizing map. Test results showed developed algorithms that speed fractal image compression using a compression algorithm, is in-creased by more than 3-fold on the basis of self-organizing artificial neural networks. The quality of the restored image after compression practically does not worsen.
References:

 

  1. Osnovy fraktalnogo szhatija izobrazhenijj. [EHlektronnyjj resurs]. Rezhim dostupa: http://habrahabr.ru/post/126653.
  2. Samoorganizujushhajasja karta Kokhonena. [EHlektronnyjj resurs]. Rezhim dostupa: https://ru.wikipedia.org/wiki/ Samoorganizujushhajasja_karta_Kokhonena.

 

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio