Abstract | Full Paper | PDF | Printer friendly format

Abstract
The recognition of objects is one of the most challenging goals in robotic vision system. The problems increase when the process of recognition involves three dimensional (3-D) objects. To overcome this problem, many researchers have proposed their own solution. In this paper, a multiview 3-D object recognition system based on neuro-fuzzy system is proposed. The system consists of two stages. The first stage is feature extraction of object. In practice, there are many techniques for feature extraction such as moment invariants, Fourier transform coefficients, M-transform etc. We use moment invariants for feature. extraction. The second stage is the process of recognition. We use Multiple Adaptive Network based Fuzzy Inference System (MANFIS) as a classifier for this stage. This method has been successfully tested using five different objects. High recognition rate was obtained using the proposed method.