search

voidlin / asmlibrary

Automatically exported from code.google.com/p/asmlibrary
0 stars 0 forks source link

Shape extraction #6

Open GoogleCodeExporter opened 8 years ago

GoogleCodeExporter commented 8 years ago 

 Hi There,

 I tried your library and it works really well. I was wondering how you would extract each shape in the model, for example, how could I get the region matching the eyes, the nose and the mouth.

 Thanks,

 Marcos

Original issue reported on code.google.com by danielg...@googlemail.com on 16 Sep 2010 at 4:49

GoogleCodeExporter commented 8 years ago 

 The fact that the "Draw" method does join every single point close to the corresponding one, means that the library is not yet ready to know what those landmarks are belonging to, hence the facial features extraction is not as simple as I thought at the beginning. 

 Is there any release coming up taking into account these difficulties?

 Thanks.

Original comment by danielg...@googlemail.com on 18 Sep 2010 at 11:12

GoogleCodeExporter commented 8 years ago 

 It seems then that this could be done by pre-processing given the landmarks are manually annotated. At the end of the day, the shape class obtained by fitting the model is the same vector as the initial one used for landmarking. While landmarking, one could annotate what positions of the vector belongs to what facial features.

 Then second part of the problem would be using a tool to see the landmarks initially so that a file with the points and their corresponding position in the vector can be created.

 I'll give a shot and let you all know about the outcomes.

Original comment by danielg...@googlemail.com on 18 Sep 2010 at 11:26