摘要
提出了基于2D图像模糊识别的3D模型检索的方法。该方法根据人识别习惯模式,从6个方向采集3D模型的2D图像。考虑到人眼对色差的敏感的特点,采用适当的均匀色空间,建立各2D图像的纹理特征。然后对这些纹理特征采用模糊C-均值算法建立模糊集,这些模糊集用于模拟人对图像模糊认识,模糊集对图像变形具有鲁棒性,故提高了提取特征的稳定性和灵活性,同时改善了检索效果与人识别效果的一致性。
A 3D model retrieval method based on 2D images fuzzy recognition is proposed According to human's habit of recognition, the method gathers the 3D model' s images from six different directions. Taking account of our eyes' sensitivity for chromatism, an appropriate even color pattern is applied. Once the textural model is established,we'll conceive a fuzzy set by FCM for the model, which simulates the recognizing outcome of human. The fuzzy set has robustness against the image distortion. The method will promote the retrieval system' s stability and flexibility. And it will improve the consistency with human recognition.
出处
《南昌大学学报(理科版)》
CAS
北大核心
2005年第5期503-507,共5页
Journal of Nanchang University(Natural Science)