Recognition of Business Objects in Street-View Images Using Sub-Space Grids

摘要

The work presented in this paper introduces sub-space grids and shows how it can be employed for recognition of business objects like bank, fire station, restaurant and store in street-view images. The paper first describes projection of segmented image area to a six dimensional feature vector space. Principal component analysis (PCA) and multiple discriminant analysis (MDA) algorithms are used to define the orientation of six feature vectors. The range of value associated with each feature vector is divided into a number of equal parts to define six dimensional sub-space grids. A recursive procedure is then used to obtain rules where sub-space grids form premises of rules. The system is tested on a dataset of 20 images (5 images of each business object type). The results show that the use of sub-spaces grids produces good results to recognize business objects in street-view images.