Fine-grained Abnormality Detection and Natural Language Description of Medical CT Images Using Large Language Models
Journal Title: International Journal of Innovative Research in Computer Science and Technology - Year 2025, Vol 13, Issue 1
Abstract
Medical report generation demands accurate abnormality detection and precise description generation from CT images. While large language models have shown promising results in natural language processing tasks, their application in medical imaging analysis faces challenges due to the complexity of fine-grained feature detection and the requirement for domain-specific knowledge. This paper presents a novel framework integrating large language models with specialized medical image processing techniques for fine-grained abnormality detection and natural language description generation. Our approach incorporates a multi-modal knowledge enhancement module and a hierarchical attention mechanism to bridge the gap between visual understanding and textual description. The framework employs an adapter-based architecture for efficient domain adaptation and introduces a medical knowledge-enhanced loss function to improve description accuracy. Experimental results on three public datasets demonstrate the effectiveness of our approach, achieving 94.6% detection accuracy and a BLEU-4 score of 0.421 for description generation, surpassing current state-of-the-art methods. The system shows particular strength in handling subtle abnormalities, with a 91.2% average precision in fine-grained detection tasks. Comprehensive ablation studies validate the contribution of each component, while qualitative analysis demonstrates the clinical relevance of generated descriptions. The proposed framework represents a significant advancement in automated medical image analysis, offering potential benefits for clinical workflow optimization and diagnostic support.
Authors and Affiliations
Zhongwen Zhou , Siwei Xia , Mengying Shu , Hong Zhou
Keywords
Related Articles
- EP ID EP753591
- DOI 10.55524/ijircst.2024.12.6.8
- Views 3
- Downloads 0
Test Cases Optimization Evaluation Using Efficient Algorithm with UML
The expenses of software testing is about 40-60% of the total cost of the software, so that reduction of test case numbers or test suite size is very much important and cannot avoid it without compromise the quality of t...
Hybrid Active Power Filter for Power Quality Improvement
A Deadbeat current controller for an LC-coupling hybrid active power filter is proposed, which can track with the reference compensation current with low steady- state error and fast dynamic response. Moreover, it can le...
Advertisement Based Lock Screen Using Location and Time Information
Advertisements are more important for locality based business to improve their sales through advertisement to attract more new customers. In this project, we integrate Advertisements template with mobile lock screen usin...
2:1 Multiplexer Design Using Lector, LCnmos, LCpmos Power Reduction Techniques with 45nm, 90nm, 180nm CMOS Technology
Today’s modern communication requires high data transmission rate and low power consumption. One of the most common concepts of data transmission can be achieved by Multiplexers. The Multiplexers are the logic designs wh...
A Study on Utilization of Waste Materials and Eco-Friendly Construction Materials to Make Green Concrete
Tyre is rubber member that provides cushion against the shocks and support load. Tyre is waste material which cause many environmental effects in all the parts of the world representing very serious threat to ecology. On...