Related Experiment Videos
Description generation of abnormal densities found in radiographs
A Abella1, J R Kender, J Starren
1AT&T Bell Laboratories, Murray Hill, NJ, USA.
Proceedings. Symposium on Computer Applications in Medical Care
|January 1, 1995
Summary
This study introduces a system that automatically describes renal stones in radiographs by analyzing spatial relationships between stones and organs. The system aims to generate descriptions consistent with those provided by radiologists, improving diagnostic accuracy.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Radiology
Background:
- Radiographic descriptions of renal stones are crucial for diagnosis and treatment planning.
- Current methods for describing renal stones can be subjective and time-consuming.
- Automating this process can enhance consistency and efficiency.
Purpose of the Study:
- To develop and present a novel system for generating radiologist-adherent descriptions of renal stones from radiographs.
- To leverage spatial relationships between renal stones and major organs for descriptive accuracy.
Main Methods:
- Image processing for precise renal stone localization.
- Inference network minimization to identify the most descriptive spatial relationships.
- Natural language generation to translate spatial data into medical terminology.
Main Results:
- The system successfully identifies renal stones within radiographs.
- It determines the most significant spatial relationships between stones and organs.
- Generated descriptions align with those produced by human radiologists.
Conclusions:
- The presented system offers an automated approach to describing renal stones in radiographs.
- This technology has the potential to standardize and improve the accuracy of radiological reports.
- Further illustration with examples will demonstrate the system's practical application.