Application of object prediction theory in object localization

Abstract

In this paper, an object localization method based on prediction theory is proposed. Prediction theory is employed to construct the network in the approach. Besides, the proposed work is applied to key component localization on a running gear. Its performance is compared with a scale-invariant feature transform (SIFT) based object localization. The proposed network is designed to represent the structure of objects, and the recognition of objects is to be accomplished after training the network. The experiment demonstrates that the proposed method can accurately localize objects in a big image by using a small amount of training data.