Title :
Training ANN for Image Discrimination Based on Wavelet
Author_Institution :
Coll. of Comput. Sci., Xi´´an Polytech. Univ., Xi´´an
Abstract :
This paper introduced a technique for training ANN for image discrimination based on Wavelet. The technique models perception of human being as process of training ANN according to visible difference predictors (VDPs). We used the Wavelet algorithm to acquire a training set, with ART-2 as representative of ANN, which is considered a tradeoff of stability-plasticity dilemma. The experiments show an accuracy rate close to 80%.
Keywords :
artificial intelligence; image processing; neural nets; prediction theory; wavelet transforms; ANN; VDP; artificial neural networks; image discrimination; stability-plasticity dilemma; visible difference predictors; wavelet method; Artificial neural networks; Computational modeling; Computer graphics; Humans; Image edge detection; Information technology; Neural networks; Pixel; Rendering (computer graphics); Stability;
Conference_Titel :
Convergence and Hybrid Information Technology, 2008. ICHIT '08. International Conference on
Conference_Location :
Daejeon
Print_ISBN :
978-0-7695-3328-5
DOI :
10.1109/ICHIT.2008.228
