LAUSR

In order to make full use of the color, shape, texture, and other features in the 3-D vehicle model, a 3-D vehicle model retrieval method is proposed based on angle structure features using render images as dataset. First, the 3-D vehicle model render images are taken as a testing set and the marked natural images are taken as a training set. The render images are classified based on their skeleton-associated shape context and the angle structure features are extracted to establish the feature library. Then, the angle structure features of the input natural image are extracted. The distance measure method is used to calculate the similarity between the angle structure feature of input natural image and those features in the feature library in order to achieve the 3-D vehicle model retrieval. The experimental results show that this retrieval method can effectively improve retrieval performance for the 3-D vehicle model.