Hello,
I am trying to put up running in my PC the Vocabulary Trees
implementation located in the vlsam stack of diamondback.
I am a bit confused when taking a look at the file
"train_weights.cpp". I am not sure what is expected as input in the
"objects.dat" file.
If I am not wrong, it reads "num_sigs" descriptors calculated for an
image (a document) from the signatures.dat file. Then, it tryies to
get "num_sigs" identifiers from the "objects.dat" file ... does this
mean that, for example for object 1, I should have "num_sigs" times
the value 1 in the "objects.dat" file?
Thank you!
(text/plain)