模式识别与人工智能
模式識彆與人工智能
모식식별여인공지능
Moshi Shibie yu Rengong Zhineng
2013年
12期
1146-1153
,共8页
图像分层表示%前馈%概率模型%潜在Dirichlet分布(LDA)
圖像分層錶示%前饋%概率模型%潛在Dirichlet分佈(LDA)
도상분층표시%전궤%개솔모형%잠재Dirichlet분포(LDA)
Image Hierarchical Representation%Feed-Forward%Probabilistic Model%Latent Dirichlet Allocation ( LDA)
现有的图像分层表示方法严格局限于前馈型方式，不能较好地解决局部模糊性等问题。基于此，文中提出一种学习和推断层次结构所有分层的概率模型，它考虑递归的概率分解过程，通过推导得到金字塔式多层结构的潜在Dirichlet分布的衍生模型。该模型存在两个重要特性：增加表示层可提高平面模型的性能；采用全Bayesian概率方法优于其前馈型实现形式。在标准识别数据集上的实验结果表明，与现有的分层表示方法相比，该模型表现出较好性能。
現有的圖像分層錶示方法嚴格跼限于前饋型方式，不能較好地解決跼部模糊性等問題。基于此，文中提齣一種學習和推斷層次結構所有分層的概率模型，它攷慮遞歸的概率分解過程，通過推導得到金字塔式多層結構的潛在Dirichlet分佈的衍生模型。該模型存在兩箇重要特性：增加錶示層可提高平麵模型的性能；採用全Bayesian概率方法優于其前饋型實現形式。在標準識彆數據集上的實驗結果錶明，與現有的分層錶示方法相比，該模型錶現齣較好性能。
현유적도상분층표시방법엄격국한우전궤형방식，불능교호지해결국부모호성등문제。기우차，문중제출일충학습화추단층차결구소유분층적개솔모형，타고필체귀적개솔분해과정，통과추도득도금자탑식다층결구적잠재Dirichlet분포적연생모형。해모형존재량개중요특성：증가표시층가제고평면모형적성능；채용전Bayesian개솔방법우우기전궤형실현형식。재표준식별수거집상적실험결과표명，여현유적분층표시방법상비，해모형표현출교호성능。
The existing image hierarchical representation methods are strict in feed-forward style, and therefore it is not able to solve problems like local ambiguities well. In this paper, a probabilistic model is proposed to learn and deduce all layers of the hierarchy together. Specifically, a recursive probabilistic decomposition process is taken into account, and a generative model based on latent Dirichlet allocation with pyramidal multilayer structure is derived. Two important properties of the proposed probabilistic model are demonstrated:adding an additional representation layer to improve the performance of the flat model and adopting a full Bayesian approach which is better than a feed-forward implementation of the model. Experimental results on a standard recognition dataset show that the proposed method outperforms the existing hierarchical approaches, and it improves the classification and the learning accuracy with better performance.
