基于模糊识别的三维模型检索  

A 3D MODEL RETRIEVAL METHOD BASED ON FUZZY RECOGNITION

作　　者：程春雷[1] 白似雪[1] 曹重华[1] 肖水晶[1] 

机构地区：[1]南昌大学计算机科学与技术系,江西南昌330029

出　　处：《南昌大学学报（理科版）》2005年第5期503-507,共5页Journal of Nanchang University(Natural Science)

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.

关 键 词：三维模型检索 2D图像 纹理 模糊识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]