WSEAS Transactions on Signal Processing
Print ISSN: 1790-5052, E-ISSN: 2224-3488
Volume 15, 2019
Implementation of Discrete Wavelet Transform on Movement Images and Recognition by Artificial Neural Network Algorithm
Authors: ,
Abstract: In this paper presented the implementation of discrete wavelet transforms (DWT) on movement image data in CCTV recordings using. Movement image on CCTV recordings is taken using background subtraction technique. Implementation of DWT on data is aimed to obtain a smaller amount of image data but not eliminating the characters of the original image characters. The application of discrete wavelet transforms is performed by filtering technique using impulse wavelet Daubechies order 4 (Db4). From the test conducted, on the first level decomposition, the data size reduction is 49.99% with the change in parameters average value of pixel is 1.19% and pixel pattern change 1.93%. In second level, the data size reduction is 24.99% with the change in parameters average value of pixel is 1.62% and pixel pattern change 2.46%. In Third level, the data size reduction is 12.48% with the change in parameters average value of pixel is 2.32% and pixel pattern change 3.84%. In fourth level, the data size reduction is 6.22% with the change in parameters average value of pixel is 2.31% and pixel pattern change 4.57% from the pattern of original image. From these results it can be concluded that wavelet transformation can be used to minimize the amount of image data without loss the characteristics of its original. In testing MLP classification by Weka 3.8, by using training set, 100% correctly classified.
Search Articles
Pages: 149-154
WSEAS Transactions on Signal Processing, ISSN / E-ISSN: 1790-5052 / 2224-3488, Volume 15, 2019, Art. #18