Identification and recognition of three dimensional (3D) objects in range data is a challenging problem. We propose a novel method to fulfill the task through two steps: 1) construct the feature signatures for the objects in the scene and the models in a 3D database; 2) based on the feature signature, find out the most similar model which decides the class of the corresponding object in the scene. We also evaluate the accuracy, robustness of the recognition method with several configurations. Our experimental results validate the effectiveness of our method.
Conference Dates: August 19-21, 2008
Conference Location: Gaithersburg, MD
Conference Title: Performance Metrics for Intelligent Systems Workshop Proceedings (PerMIS)
Pub Type: Conferences
accuracy evaluation, object recognition