Hi,I'm now working on locating on a specific scene. And I think training a offline BoW for this scene can improve accuracy. How did you generate the dictionary brief_k10L6.bin? it uses DBOW2? Can you tell me some details?
Thanks a lot !!!
I think so. It should be using DBow2 which uses DLib. k=10 is the branching factor and L=6 is the depth factor. There is a demo in DLoopDetector and you can generate your own voc file.
Hi,I'm now working on locating on a specific scene. And I think training a offline BoW for this scene can improve accuracy. How did you generate the dictionary brief_k10L6.bin? it uses DBOW2? Can you tell me some details? Thanks a lot !!!