Further study on camera position estimation from image by ANFIS

It is clear that different images will be obtained when one takes pictures from different camera positions. One can easily determine the characteristics of the captured image by projective geometry. However, it is hard to estimate the camera position from only one image. Machine learning methods are very useful for the identification of nonlinear relations. In this research, comparisons of the known situations of images and earlier works are considered. An adaptive neuro-fuzzy inference system (ANFIS) network design is deployed and used for estimating the camera position. From the experimental results, it is evident that the proposed method can estimate the center of the camera correctly and effectively.