log.txt~

27 April 18
-Trying to figure out data abstraction for C++, specifically, trying to get an idea for how to structure the massive project that is sort of coming together.
I currently have the idea for a LightField Class, but I have a number of other classes/methods that need to go into it without necessarily rewriting all of those.
More importantly though, I need to get this structure from motion stuff working so I can get the depth. Once I get the depth, I'll still need to add it to the 
DL program. *Also need to look at the SIGGRAPH documentation and figure out what to start writing for that!
Something along the lines of creating a new dataset, using deep learning on it, getting depth from it.
So, two big things for the next few days:
Start SIGGRAPH Paper
LightField Class
-------------------------------------------------------------
LightField Class Structure


SFMModule	GeneralRender		ImageRenderer		GetAndTriangulate
private
-Match		-			-			-		
-CheckRot	-			-			-	
-FindCamMat	-			-			-	
-Triang		-			-			-	
-		-			-DrawAperture		-	
public
-SavePCD	-			-CBinMain!		-GetImages	
-		-			-			-Triangulate	
-		-			-			-Normalize	
-		-			-			-CalculateArea	
LightField Class:
private
-LoadImages
-NormalizeVector
-Load Poses

public
-TriangulateLf

-PercentOfLF
-PointsInFace
-FaceVertices

-DepthMultiView //uses camera poses
-DepthFromDefocus
-DepthFromSfm
-ViewByRays
-GeneralRender

Will also need to add visualizer for point clouds
Note that there are a number of functions hidden in the iccp tests etc