Fine-grained Abnormality Detection and Natural Language Description of Medical CT Images Using Large Language Models

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

Research on Decryption Methodologies and Key Aggregate Searchable Encryption for Data Security Storage in Cloud

Numerous firm’s architectures management of data guarantees substantially alter method, gain access to maintain private commercial business. Occur additional facts protection problems. Existing statistics safety methods...

Managing Object Oriented Software Understandability: A Design Perspective

Estimating understandability of object oriented software early in the development process; particularly at design phase greatly reduce the overall development cost and effort. To design and deliver quality products insid...

Soil Stabilization Using Crumb Rubber Powder

Because there are more used car tyres produced each year, disposing of them has become a significant environmental issue on a global scale. Utilizing used tyres will reduce the effect on the environment and increase reso...

Efficient Secured Two Party Computing with Encrypted Data for Public Cloud

In this project we intervened certificateless encryption plan without matching operations for safely imparting public cloud. Intervened certificateless open key encryption (mCL-PKE –Mediated certificateless public key en...

Version Locking Mechanism in Database

The distributed database provides a resource sharing environment, where multiple transactions at different sites coexist in order to access the resources. In this paper we investigate multi version locking protocol in di...

Download PDF file
  • EP ID EP753591
  • DOI 10.55524/ijircst.2024.12.6.8
  • Views 42
  • Downloads 0

How To Cite

Zhongwen Zhou, Siwei Xia, Mengying Shu, Hong Zhou (2025). Fine-grained Abnormality Detection and Natural Language Description of Medical CT Images Using Large Language Models. International Journal of Innovative Research in Computer Science and Technology, 13(1), -. https://europub.co.uk/articles/-A-753591