A comparison of the animal recognition between the real objects and the modeled 3D objects

This paper proposed about comparison of the animal recognition processed on the real world objects and the modeled 3D objects. The object classification consists of the three main steps - feature extraction, training classifier and image evaluation. Feature description is based on locale visual descriptors like SIFT or SURF. Support Vector Machine (SVM) in combination with the bags of visual keypoints (BOW) is used to classify descriptors. The 3D models were created using photogrammetry and 3D modeling for ideal 3D models of mammals. This method of 3D model creation is described in details.