Autor(es)

Ribas, Lucas C. ; Bruno, Odemir M.

• Tipo de documento: Artigo Científico
• Fonte: PROGRESS IN PATTERN RECOGNITION, IMAGE ANALYSIS, COMPUTER VISION, AND APPLICATIONS, CIARP 2018; v. 11401, p. 8-pg., 2019-01-01.

Resumo

Shape is an important characteristic used by different classification tasks in computer vision. In particular, shape is useful in many biological problems (e.g. plant species recognition and fish otolith classification), which are challenging due to the diversity found in nature. This paper proposes a novel method for shape analysis and classification based on deterministic partially self-avoiding walks (DPSWs) on networks. First, a shape contour is modeled as a network by mapping each contour pixel as a vertex. Then, deterministic partially self-avoiding walks are performed on the network and a robust shape signature is obtained using statistics of the trajectories of the DPSWs. We evaluate this feature vector in a classification experiment using two different natural shape databases: USPLeaves and Otolith. The experimental results demonstrate a high classification accuracy of the method when compared to the other methods. This suggests that our method is a promising option for the classification task in biological problems. (AU)

• Processo FAPESP: 16/18809-9 - Deep learning e redes complexas aplicados em visão computacional
• Beneficiário: Odemir Martinez Bruno
• Modalidade de apoio: Auxílio à Pesquisa - Parceria para Inovação Tecnológica - PITE
• Processo FAPESP: 16/23763-8 - Modelagem e análise de redes complexas para visão computacional
• Beneficiário: Lucas Correia Ribas
• Modalidade de apoio: Bolsas no Brasil - Doutorado
• Processo FAPESP: 14/08026-1 - Visão artificial e reconhecimento de padrões aplicados em plasticidade vegetal
• Beneficiário: Odemir Martinez Bruno
• Modalidade de apoio: Auxílio à Pesquisa - Regular