A new algorithm for infrared image restoration based on multi-scale morphological wavelet and Hopfield neural network

Abstract

Based on the complexity and randomness of the infrared image degradation factors, and integrates the strong de-noising features of multi-scale morphological wavelet and the salient problem solving features of Hopfield neural network in optimization, this paper presents a new algorithm for infrared degraded image restoration. The algorithm takes advantage of the continuous recycle between "multi-scale morphological wavelet de-noising" and "Hopfield neural network iteration" so as to makes access to a better recovery of infrared images. The algorithm also solves the problems in noise suppression and image detail protection of traditional Hopfield neural network image restoration algorithm and successfully protects the edge of the recovery images and details. Simulation results prove the effectiveness of the recovery algorithm.

QR Denoising using a Hopfield Network

“…A previous work uses the super-resolution technique that generates a high -resolution image from multiple low -resolution images to improve the performance of low -resolution QR code recognition. [4,5,6,7] In addition, some researchers use various binarizations to improve the non -uniform background and uneven light problems. Researchers propose several different methods for locating and extracting QR image code in the extraction of the QR code.…”

Equality in Soft Erosion, Soft Dilation and Soft Open in Multi Scale Environment.

“…JIAN-HUI TAN etc. proposed (25) a new type of process using MSMM for smoothening of infrared imagery. It will have complexity due to noise.…”

WNN in Image Restoration for Industrial Applications

“…Based on the complexity and randomness of the infrared image degradation factors, and integrates the strong de-noising features of multi-scale morphological wavelet and the salient problem solving features of Hopfield neural network in optimization. In this Ao-Chang Pan et al [4] presented a new algorithm for infrared degraded image restoration. Algorithm takes advantage of continuous recycle between multi-scale morphological wavelet denoising and Hopfield neural network iteration so as to makes access to a better recovery of infrared images.…”