Evaluating LICF for indoor modeling

In this paper, the suitability of features for the modeling of artificially constructed indoor environment was verified by comparing LICF with other features. The repeatability, accuracy and Invariance were measured for the assessment. In the same indoor place, images were obtained by making diverse environmental changes. The same image was used for the feature extraction process, and the extracted features were reconstructed in the space via manual matching to ensure the independency of the matching technique for the stereo image during accuracy measurement. The performances of feature for the environment modeling were examined by comparing the distance between features in actual objects or environment.