Skip to content

A study on image captioning for the fashion domain

Notifications You must be signed in to change notification settings

NoLogicPlease/Visionizer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 

Repository files navigation

Visionizer

DOI

Platforms for fashion e-commerce are rising in popularity. However, the majority of the manual work still goes into scanning, rendering, and captioning fashion products. The task of creating a textual description of a fashion item from a photograph depicting it is covered in this research on fashion image captioning. For this work, we create the FashionCap dataset, which consists of 40,000 captions for over 290,000 photos and it is based on the original counterpart FashionGen. We conduct a thorough analysis using several neural architectures on three fashion captioning datasets.

Datasets

The following datasets were used, and are available here: https://doi.org/10.5281/zenodo.7196078

  • FashionCap
  • ReducedFACAD
  • ReducedInFashAI

Dependencies

  • tensorflow
  • numpy
  • pandas
  • nltk
  • bert_score
  • matplotlib
  • tqdm
  • numba

Best BLEU score instance

Worst BLEU score instance

Attention Weights during generation

Releases

No releases published

Packages

No packages published