采用累计贡献率的方法，匹配追踪和正交匹配追踪，加入重复控制，isodata 迭代自组织的数据分析，多姿态，多角度，有不同光照,基于kaiser窗的双谱线插值FFT谐波分析，一种流形学习算法（很好用），基于欧几里得距离的聚类分析。- The method of cumulative contribution rate Matching Pursuit and orthogonal matching pursuit, Join repetitive control, Isodata iterative self-organizing data analysis, Much posture, multi-angle, have different light, Dual-line interpolation FFT harmonic analysis kaiser windows, A fluid manifold learning algorithm (good use), Clustering analysis based on Euclidean distance.