Related Experiment Video
Updated: Oct 7, 2025
04:23
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
2.0K
Attention based automated radiology report generation using CNN and LSTM.
Mehreen Sirshar1, Muhammad Faheem Khalil Paracha1, Muhammad Usman Akram1
1Department of Computer and Software Engineering, College of Electrical and Mechanical, National University of Sciences and Technology, Islamabad, Pakistan.
Plos One
|January 6, 2022
Summary
This study introduces an advanced AI model for generating radiology reports from X-rays, improving disease diagnosis accuracy. The novel approach enhances clinical decision-making by reducing the burden on radiologists.
Area of Science:
- Artificial Intelligence
- Medical Imaging Analysis
- Natural Language Processing
Background:
- Automated radiology report generation holds significant potential for improving clinical diagnosis.
- Current methods for chest X-ray (CXR) analysis lack sufficient accuracy, hindering comprehensive medical report creation.
- Hybrid approaches combining computer vision and natural language processing are emerging for auto medical report generation.
Purpose of the Study:
- To develop a novel approach for accurate automated generation of radiology reports from medical images.
- To address the limitations in sensitivity and accuracy of existing techniques for CXR findings interpretation.
- To reduce the workload of radiologists and assist in manual report writing.
Main Methods:
- A hybrid model integrating Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) for disease detection.
- Utilizing an attention mechanism for sequence generation based on detected diseases.
- Employing Indiana University CXR and MIMIC-CXR datasets for model training and validation.
Main Results:
- The proposed model achieved state-of-the-art performance compared to baseline solutions.
- Demonstrated improved accuracy in detecting diseases from chest X-ray images.
- Quantitative evaluation using BLEU-1, BLEU-2, BLEU-3, and BLEU-4 metrics confirmed the model's effectiveness.
Conclusions:
- The novel hybrid AI model significantly enhances the accuracy of automated radiology report generation.
- This approach offers a promising solution for improving diagnostic efficiency and reducing radiologist workload.
- The model's performance on benchmark datasets indicates its potential for clinical application in medical imaging analysis.