January 2018
Intermediate to advanced
310 pages
7h 48m
English
DeepRank proposed by Wang et al. (https://users.eecs.northwestern.edu/~jwa368/pdfs/deep_ranking.pdf) is used to rank images based on similarity. Images are passed through different models as shown here:

The triplet loss is computed here as well and backpropagation is done more smoothly. Then the image can be converted to a linear embedding for ranking purposes, as shown:

This algorithm is highly useful for ranking purposes.
Read now
Unlock full access