Self-organization for object extraction using a multilayer neural network and fuzziness measures

Abstract

Abstract-The feedforward multilayer perceptron (MLP) with back-propagation of error is described. Since use of this network requires a set of labeled input-output, as such it cannot be used for segmentation of images when only one image is available. (However, if images to be processed are of similar nature, one can use a set of known images for learning and then use the network for processing of other images.) A self-organizing multilayer neural network architecture suitable for image processing is proposed. The proposed architecture is also a feedforward one with back-propagation of errors; but like MLP it does not require any supervised learning. Each neuron is connected to the corresponding neuron in the previous layer and the set of neighbors of that neuron. The output status of neurons in the output layer is described as a fuzzy set. A fuzziness measure of this fuzzy set is used as a measure of error in the system (instability of the network). Learning rates for various measures of fuzziness have been theoretically and experimentally studied. An application of the proposed network in object extraction from noisy scenes is also demonstrated.

A Comparative Study on the Development of Binary Object Extraction System using Different Self Organizing Neural Network

“…The result is shown in Fig 4 and Fig.5 below. The percentage of correct classification of pixels (pcc) [22] is calculated on the basis of the achieved image and the original image to determine the quality of the extracted image. It uses the formula: pcc = * 100, where t cc stands for total number of correctly classified pixels and the total number of pixels in the image scene is designated to t np …”

“…The results show the restoring capabilities of the network in the extracted images. The main problems with these architectures lie in the adjustment and reassignment of the interconnection weight [22] matrices required for the processing of the information. Most of these approaches use back-propagation algorithm for updating the interconnection weights.…”

Localization of tissues in high-resolution digital anatomic pathology images

“…[13][14][15][16] Unsupervised NN was also used for segmentation problem to reduce the limitation of supervised learning such as Kohonens learning vector quantization (LVQ) network. 17 Hall et al, 13 perfomed a comparison between a Neural Network and fuzzy clustering techniques in segmenting MR Images of the Brain.…”

Multipurpose image recognition based on active search and adaptive fuzzy inference neural network

“…Rule-based systems have had limitations due to the necessity of specific design for each purpose, and there have been intensive studies of image recognition utilizing the learning functions of neural networks [12][13][14][15]. The authors have considered a fuzzy inference neural network (FINN) [16] in which the learning power of the neural network and the advantages of fuzzy inference with knowledge processing function are combined.…”