原文：http://www.jb51.cc/article/p-fwhgxfxm-bcb.html

从该blog开始，逐步介绍DPM + Latent SVM。关于OpenCV下DPM+Latent SVM简单介绍参考上一篇博文：OpenCV Latent SVM Discriminatively Trained Part Based Models for Object Detection

以Cat.xml（opencv安装目录下sample/c内）为例

<Model>

<!-- Number of components -->

<NumComponents>2</NumComponents>

<!-- Number of features -->

<P>31</P>

<!-- score threshold -->

<scoreThreshold>-1.0028649999999999</scoreThreshold>

<Component>

<!-- Root filterdescription -->

<RootFilter>

<!-- Dimensions-->

<sizeX>5</sizeX>

<sizeY>9</sizeY>

<!-- Weights(binary representation) -->

<Weights>

...

</Weights>

<!-- Linear termin score function -->

<LinearTerm>-2.2535784347835031</LinearTerm>

</RootFilter>

<!-- Part filtersdescription -->

<PartFilters>

<NumPartFilters>6</NumPartFilters>

<PartFilter>...</PartFilter>

</PartFilters>

</Component>

<LinearTerm>-2.5835343890077622</LinearTerm>

<!--Part filters description -->

<!-- Part filter? description -->

</Model>

PartFilter内部结构：

<PartFilter>

<sizeX>6</sizeX>

<sizeY>6</sizeY>

<!-- Weights (binary representation)-->

<Weights></Weights>

<!-- Part filter offset -->

<V>

<Vx>3</Vx>

<Vy>1</Vy>

</V>

<!-- Quadratic penalty functioncoefficients -->

<Penalty>

<dx>0.0004031731821276</dx>

<dy>-0.0003745111759062</dy>

<dxx>0.0100010270581015</dxx>

<dyy>0.0205820897831230</dyy>

</Penalty>

</PartFilter>

注：1.里面的数据仅是为了说明，weight节点数据量太大，省略，PartFilter节点有点复杂，单列出来；

2.其中有些简洁说明性文字，不是很全，PartFilter中Vx，Vy为HOG的位移。Penalty节点下的各项是为了计算deformation model。