ASM中基于不变矩图形畸变主动检测与修正

主动形体模型是可变形模板技术中一种重要的方法，在医学图像分析和机器视觉中得到越来越广泛的应用。但是该方法在图像的匹配过程中，如果目标图像不够清晰或者模型初始位置不理想，会产生畸变的匹配结果。而且该模型没有一个有效的对畸变图形进行检测和修正的策略。提出了一种基于边界矩不变量的主动检测和修正方法，通过引入边界矩不变量，对模型的形体变化进行量化，并根据从训练集中获取的统计信息，对变形过程中的模型形体的变形进行检测和修正。实验结果表明，该方法能够在很大程度上解决匹配过程中的畸变问题，并且相比传统主动形体模型，所消耗的时间增加很少，对算法的效率并不影响。

Active Detection and Correction of Shape Distortions Based on Moment Invariants for Active Shape Models

The active shape model is one of the important method in the field of the deformable models. In the fields of the medical image analysis and machine vision, it is an ever-increasingly broad application method. During the process of image fitting, however, distortions and displacements often happen if the target is not clear or the initial position of the model is not ideal, and there is a lack of effective correction strategies. In this paper, we firstly propose an active detection and correction based on the boundary moment invariants. By introducing the boundary moment invariants, the changes of the model can be quantification, and can be detected and corrected according to the statistical information which is obtained from the training set. Using the proposed method, distortions are effectively avoided and the accuracy of fitting result is obviously increased with little extra time. Finally, the results of our practical implementation prove that the proposed strategy works satisfactorily.