Proposta de um novo descritor de imagens baseado em autômatos celulares

Abstract

The feature extraction process from images is fundamental in solving several computational problems. This research presents a new image analysis and description approach based on cellular automata. This approach allows modeling the discrete local relationship between pixels, considering shape, intensity, and relationship between objects in the image. Based on Conway’s Game of Life, we created four methods to analyze the relationship between pixels in an image and extract their features. In addition, we proposed different combinations of the results of each method, which allowed us to obtain better results during the experiments. To validate the proposed approach, we used public databases to compare our results with the results of the state-of-the-art feature extraction approaches. We used accuracy to measure the performance of the methods. After the experiments, the proposed approach presented an accuracy of up to 83.90% and 73.30% for Corel-1K and Pap-smear datasets, respectively. The results show that the proposed approach outperforms state-of-the-art approaches and can be successfully applied in image feature extraction.