Skip to content

An approach for solving the problem of medical visual question answering

Notifications You must be signed in to change notification settings

uakarsh/med-vqa

Repository files navigation

About the Project:

Problem : Question Answering for the Medical Images

Implementation: I have taken inspiration from Stacked Attention Network, and the slides as mentioned above, and implemented in Tensorflow 2.0, however, I have made some changes, because I did not understand those things, will improve in the near future.

Setup:

GOOGLE COLAB: What you need to do is:

Download :

    1. trainset.json
    2. testset.json
    3. VQA Image Folder
    4. Cache Folder  (contains the pickle file, for converting the answers to labels, and vice versa, and the mapping for dictionary and answer)

Upload these projects to your google drive, and then follow the instructions that are present in the VQA.ipynb notebook.

Results:

alt text alt text

📺 Some of my projects:

âš¡ GitHub Stats

Anurag's github stats

About

An approach for solving the problem of medical visual question answering

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published