Open AccessA NOVEL ALGORITHM BASED ON CASCADING OF NEURAL NETWORK MODELS AND WAVELET TRANSFORM FOR IMAGE ENHANCEMENT
Author(s) -
Sandeep Verma,
Hitesh Gupta
Publication year - 2013
Publication title -
graduate research in engineering and technology
Language(s) - English
Resource type - Journals
ISSN - 2320-6632
DOI - 10.47893/gret.2013.1024
Subject(s) - artificial intelligence , histogram equalization , wavelet transform , computer science , pattern recognition (psychology) , wavelet , adaptive histogram equalization , thresholding , stationary wavelet transform , computer vision , balanced histogram thresholding , histogram matching , discrete wavelet transform , image processing , histogram , image (mathematics)
Image enhancement and restoration is pre-request of computer vision. The distortion and degradation of image suffered the process of pattern matching and quality of image. Wavelet is very important transform function play a role in image enhancement and image de-noising. The concept of wavelet used as soft thresholding and hard thresholding. A processing of data through wavelet is very efficient in process of neural network. In this paper we discuss the proposed algorithm for image enhancement based on self organized map network and wavelet transform. Basically self organized map network is unsupervised training mechanisms of pattern, due to this reason the processing of network is very fast in compression of another artificial neural network method. And the combination of wavelet and self organized map network have great advantage over conventional method such as histogram equalization and multi-point histogram equalization and another conventional technique of image enhancement.