README 

Directory Info
=============================
1. src/
	This directory contains the main source code.  


2. dramanan_files/
	This directory contains files which are originally downloaded from the source link:
"http://www.ics.uci.edu/~xzhu/face/face-release1.0-basic.zip". Some of the files are modified
according to our need.


3. opencv_files/
	This directory contains the OpenCV viola-jones face detector executable alongwith
modified haarcascade xml files.


4. out_dir/
	The output images and the viola-jones detection files for each image are directed to
this folder. After the code is done running, get the output images from this folder.


5. images/
	This folder contains sample images obtained from our driver dataset and some other
random images. This folder also has the sample "list.txt" file.


Running the code
=============================
The main folder has the "demo.m" file.
To run the code. Follow these steps:
1. Extract the tar.gz file.
2. Start matlab.
3. Move to home folder i.e near_realtime_faceparsing_v1.0/
4. If you want to run the code on:
   i. Image set, then put the images inside a directory and get the path of this directory.
   ii. Video. Extract the video frames and put them in a directory and get the of the same directory.
   As an example check the "near_realtime_faceparsing_v1.0/images" directory. 
   NOTE: Images should be in "jpg" format.
5. Make a txt file and list out the image names in the order you want them to be processed.
   As an example check "near_realtime_faceparsing_v1.0/images/list.txt" file.
6. Run "demo" with the following parameter:
   i.   global_trflag : 0 if you don't need tracking. 1 if tracking is required.
   ii.  in_dir : This is the absolute path of the input image directories.
   iii. in_file_list : The path to the file listing the image names sequentially.

   As an example :
   >> demo(1,'images/','images/list.txt');