VQA is a multidisciplinary problem which combines two modalities: text and image. It requires computer vision and NLP techniques (probably, reasoning techniques too). The task is to answer a question correctly, where the question is accompanied by an image on which it is based.
This assignment is a domain specific challenge. ImageCLEF defines the challenge as follows: “Given a medical image accompanied with a clinically relevant question, participating systems are tasked with answering the question based on the visual image content”.
Please refer to the image below for Baseline Model Architecture:
For more details related to implementation and results of the project, please refer to file AMP REPORT.pdf and for details regarding code, pease refer to file ReadMe.txt in Code Folder.
Contributors: Meghana Kotagiri (meghanakotagiri) G Neha Reddy (nehareddyg)