Classificação e recuperação de imagens por cor utilizando técnicas de inteligência artificial

AUTOR(ES)

Túlio Cléber Bender

DATA DE PUBLICAÇÃO

2003

RESUMO

Image retrieval and classification are today the subject of extensive research. This topic poses both theoretical and practical challenges as researchers attempt to give machines such as computers and robots the ability to see. Image retrieval and classification are part of a wider field known as Computer Vision, which encompasses several practical applications such as image retrieval from databases storing only raw images, biometric recognition (from images of finger-prints, face or iris), retrieval of visual trademarks and logos from advertisements, location of objects in a scene and vision techniques in robotics. The research developed in this work is focused on obtaining a generalization of characteristics extracted from a collection of images belonging to a single class using supervised learning techniques. The result is a model that identifies a given class of images. To achieve this, a review of the state-of-the-art in content-based image retrieval systems and Machine Learning techniques was needed. There is also a discussion on the separation surfaces that an artificial neural network draws during the training stage and how they affect retrieval performance. A thorough study on the geometric understanding of these separations surfaces in the multidimensional input space is made and two iterative techniques to help achieve closed separation surfaces are proposed. This work resulted in the development of a software package (I Match) that performs image retrieval and classification using Machine Learning techniques based on color. A choice was made for Artificial Neural Networks using Multilayer Perceptron architecture with the Cascade Correlation learning algorithm to accomplish the task. Unlike traditional methods, which rely on complex techniques of image pre-processing and matching algorithms, I Match takes advantage of supervised learning to learn the model of a collection of patterns representing the pattern which should retrieved from a collection of images. The use of I Match demonstrated the validity, the advantages and the issues that require further study of the approaches proposed in this research. Three image databases were used to validate the two approaches proposed to achieve closed separation surfaces, which made it possible to characterize the behavior of the systems using artificial pictures, pictures taken under controlled conditions and pictures taken without any kind of conditioning.

ASSUNTO(S)

superfícies de separação fechadas recuperação de imagens imagens; processamento digital; classificação; recuperação; rede neural artificial; base de dados; pesquisa; armazenamento redes neurais artificiais color ciencia da computacao artificial neural networks closed separation surfaces classificação de imagens cor image retrieval image classification

Documentos Relacionados

• Estimativa do Afilamento do Fuste de Araucária Utilizando Técnicas de Inteligência Artificial
• Análise e classificação de imagens de lesões da pele por atributos de cor, forma e textura utilizando máquina de vetor de suporte
• Análise de tendência do índice Bovespa em tempo real utilizando técnicas de inteligência artificial
• Formação de grupos colaborativos em cursos a distância via web utilizando técnicas de inteligência artificial
• Color-based image retrieval using discrete distribution features analysis