Image Captioning App

About: Used Merge model developed by Tanti et al. in 2017 as a refereance for the base model for image captioning

Dataset: Acquired dataset from kaggle : Flicker8k

• Here i used Tensorflow tokenizers to get a word dictionary for encoding and de-coding for sequence processing.

• Used Tensorflow library for preprocessing the text data.

• For giving the image input to the model, I used various different architecture's for transfer learning.

• Got the best loss by using VGG16. Rather then using cool architecture focused to solve the business problem with the best solution possible.

The model used for input of the processed data as: