There are fascinating advancements in the field of machine learning - the composition of image descriptions can be done with recurrent neural nets that have memory. There is an amazing introduction by Andrej Karpathy: http://karpathy.github.io/2015/05/21/rnn-effectiveness/
You are viewing a single comment's thread from:
mcsvi, Oh, yeah, I meant this work of Li Fei-Fei and Andrej Karpathy "DenseCap: Fully Convolutional Localization Networks for Dense Captioning". Thank you for the great link!