Insik Kim
Insik Kim
We make visual words with various configuration for either experiment purpose or optimization for dataset/task. Handling this configuration in filename is very sloppy way. Make separate config fie. Config file...
For PQ, we have to train encoder (give PQ code for a given original feature) first. If we use too small number of examples for training encoder, the encoder is...
Intuitively, too small vocab size give less discriminate power, but we can build the vocabulary faster. * i.e. k-means clustering with smaller k moderate size of vocabulary gives compact representation...