Saved in:
Bibliographic Details
Main Authors: Safwan, Itbaan, Shaikh, Muhammad Annas, Haaris, Muhammad, Khan, Ramail, Tahir, Muhammad Atif
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.04384
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911251604439040
author Safwan, Itbaan
Shaikh, Muhammad Annas
Haaris, Muhammad
Khan, Ramail
Tahir, Muhammad Atif
author_facet Safwan, Itbaan
Shaikh, Muhammad Annas
Haaris, Muhammad
Khan, Ramail
Tahir, Muhammad Atif
contents We present a multi-task framework for the MediaEval Medico 2025 challenge, leveraging a LoRA-tuned Florence-2 model for simultaneous visual question answering (VQA), explanation generation, and visual grounding. The proposed system integrates three curated datasets: (1) Kvasir-VQA-x1 for question-answer learning, (2) a synthetically enriched explanation dataset offering structured medical reasoning, and (3) text-to-region pairs linking visual features with segmentation masks. This multi-task setup enables the model to jointly learn visual grounding, reasoning, and interpretation, producing responses that are both accurate and interpretable. Extensive evaluation demonstrates that our approach substantially improves over single-task baselines in both answer accuracy and visual localization, highlighting the effectiveness of grounded multi-task learning for medical VQA applications.
format Preprint
id arxiv_https___arxiv_org_abs_2511_04384
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-Task Learning for Visually Grounded Reasoning in Gastrointestinal VQA
Safwan, Itbaan
Shaikh, Muhammad Annas
Haaris, Muhammad
Khan, Ramail
Tahir, Muhammad Atif
Computer Vision and Pattern Recognition
Machine Learning
We present a multi-task framework for the MediaEval Medico 2025 challenge, leveraging a LoRA-tuned Florence-2 model for simultaneous visual question answering (VQA), explanation generation, and visual grounding. The proposed system integrates three curated datasets: (1) Kvasir-VQA-x1 for question-answer learning, (2) a synthetically enriched explanation dataset offering structured medical reasoning, and (3) text-to-region pairs linking visual features with segmentation masks. This multi-task setup enables the model to jointly learn visual grounding, reasoning, and interpretation, producing responses that are both accurate and interpretable. Extensive evaluation demonstrates that our approach substantially improves over single-task baselines in both answer accuracy and visual localization, highlighting the effectiveness of grounded multi-task learning for medical VQA applications.
title Multi-Task Learning for Visually Grounded Reasoning in Gastrointestinal VQA
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2511.04384