observations.log

27th Aug, 2021

##################################################
about files:
_norm files -> files in which point cloud is normalized into unit sphere
_out files -> files infered by the learned autoencoder model

##################################################
experiment-1: 
Ran the autoencode with the model of fxia on two classes (screw and knife) for epoch of 200
Results: F:\projects\ai\autoencoder\unseen_data\results\experiment-1
Observation: 
the outer shape is captured but not accurately
cannot make out the shape if its a screw or knife

##################################################
experiment-2:
Ran the autoencoder with the model of fxia only on knife dataset for epoch of 200
Results: F:\projects\ai\autoencoder\unseen_data\results\experiment-2
observation:
Results doesn't seem to reproduce the excat shape of the knife, but its better then experiment-1 result

##################################################
experiment-3: