Abstract. This paper describes a novel neural network based multiscale image restoration approach. The method uses a Multilayer Perceptron (MLP) trained with synthetic gray level images of artificially degraded co-centered circles. The main difference of the present approach to existing ones relies on the fact that the space relations are used and they are taken from different scales, which makes it possible for the neural network to establish space relations among the considered pixels in the image. This approach attempts at coming up with a simple method that leads to an optimum solution to the problem without the need to establish a priori knowledge of existing noise in the images. The multiscale data is acquired by considering different window sizes around a pixel. The performance of the proposed approach is close to existing restoration techniques but it was observed that the resulting images showed a slight increase in contrast and brightness. The proposed technique is also used as a preprocessing phase in a real-life classification problem of medical Magnetic Resonance Images (MRI) by using a fuzzy classification technique.

No abstract

This paper describes a neural network based multiscale image restoration approach in which multilayer perceptrons are trained with artificial images of degraded gray level cocentered circles. The main objective of this approach is to make the neural network learn inherent space relations of the degraded pixels in the restoration of the image. In the conducted experiment, the degradation is simulated by submitting the image to a low pass Gaussian filter and the addition of noise to the pixels at pre-established rates. The degraded image pixels make the input and the non-degraded image pixels make the output for the supervised learning process. The neural network performs an inverse operation by recovering a quasi non-degraded image in terms of least squared. The main difference of the approach to existing ones relies on the fact that the space relations are taken from different scales, thus providing relational space data to the neural network. The approach is an attempt to develop a simple method that may lead to a good restored version of the image, without the need of a priori knowledge of the possible degradation cause. Considering different window sizes around a pixel simulates the multiscale operation. In the generalization phase the neural network is exposed to indoor, outdoor, and satellite degraded images following the same steps use for the artificial circle image. The neural network restoration results show the proposed approach performs similarly to existing methods with the advantage it does not require a priori knowledge of the degradation causes.

This paper presents a new approach to image restoration based on ANN, considering the learning of the inverse process using a standard image for training under a multiscale approach. Dierent models of ANN were tested and compared with the traditional techniques. The standard image was articially degraded to simulate some types of frequent degradation problems. Due to the huge amount of data generated for training the ANN, this paper uses clustering techniques to reduce the training set. The paper proposes a simple restoration method that leads to a sub-optimal solution without the need of prior knowledge estimation of the degradation phenomenon. The ANN based lters were tested with dierent kinds of degraded images. The mean squared error and the signal-to-noise ratio were used as performance indices to measure the quality of the results of the ANN and of some of the existing methods for comparison. The results show that the ANN based restoration algorithms as proposed in this paper are eective restoration methods. The main advantage of the proposed approach is related to the fact that it does not require an estimation of prior knowledge of the degradation causes for each image.

No abstract

No abstract

No abstract