Image caption generation

In the previous case study, we learned about using CNN, RNN, and CTC loss together to transcribe the handwritten digits.

In this case study, we will learn about integrating CNN and RNN architectures to generate captions for a given picture.

Here is a sample of the picture:

• A girl in red dress with Christmas tree in background
• A girl is showing the Christmas tree
• A girl is playing in the park
• A girl is celebrating Christmas

Getting ready

In this section, let's list the strategy that we shall adopt to transcribe pictures:

• We'll be working toward generating captions from pictures by working on a dataset that has images as well as the captions associated with the images. Links of datasets...