Results different from "Face Detection, Pose Estimation and Landmark Localization in the Wild"

[keithzhou]

I am trying to use the modelTransfer to transfer the model from "Face Detection, Pose Estimation and Landmark Localization in the Wild"(http://www.ics.uci.edu/~xzhu/face/) to PartsBasedDetector. However, it seems the it misses most of the faces. I am wondering:
-> is this algorithm equivalent(roughly,within in rounding error etc) to Zhu's mixtures of trees algorithm(after transformed the model)?

Any help is appreciated