deletions | additions       diff --git a/After_preprocessing_four_facial_local__.tex b/After_preprocessing_four_facial_local__.tex  index 12d4ec3..337b05b 100644  --- a/After_preprocessing_four_facial_local__.tex  +++ b/After_preprocessing_four_facial_local__.tex 
...
After preprocessing, four facial local image discriptors are extracted from the depth and RGB images. In contrast to global face discriptors which compute features directly from the entire face image, local face discriptors representing the features in small local image patches have shown to be more effective in real world condition. The four local discriptors are Local Binary Patterns (LBP), is a very popular texture descriptor. It is derived from a general definition of texture in a local neighborhood. The discriminative power, computational simplicity and tolerance against monotonic gray scale changes are behind the great success of LBP in many computer vision problems. Local Phase Quantization (LPQ), an image is described using the phase information of short term Fourier transform (STFT) locally computed on a rectangular window at each pixel. Histogram of Oriented Gradients (HOG), It describes local object appearance and shape within an image by the distribution of intensity gradients or edge directions. The magnitudes of the gradient at each pixel are accumulated in to a histogram according to the gradient direction. Last is Binarized Statistical Image Features (BSIF) is to automatically learn a fixed set of filters from a small set of natural images. The set of filters are learnt based on statistics of training images \citet{kannala2012bsif}. Fig 3 17  is an example of result when applying the four local discriptors on face texture and depth images acquired with Kinect sensor for a subject from the FaceWarehouse database. \citet{boutellaa2015use}