Image pattern recognition by edge detection using discrete wavelet transforms

Edge is the high-frequency part of an image and represents location where abrupt change takes place in intensity luminescence. detection basic step feature extraction pattern recognition any image. Wavelet transforms extract low information signal separately. In two-dimensional wavelet transforms, decomposed into four sub-images one approximation three different images (horizontal, vertical, diagonal images) each decomposition level. The differences’ show how neighboring pixels differ horizontal, vertical directions. coefficients are forced to zero inverse transformed. As reconstructed shows edges describes its pattern. Using Haar at level 1, 2 3, by edge performed discussed.