Related Experiment Videos
Skeletal growth estimation using radiographic image processing and analysis
S Mahmoodi1, B S Sharif, E G Chester
1Department of Electrical and Electronic Engineering, University of Newcastle, Newcastle upon Tyne, UK.
Summary
This study introduces an automated vision system for estimating children's skeletal growth from hand radiographs. The system achieves high accuracy, comparable to expert radiologists, showing potential for pediatric medicine applications.
Area of Science:
- Medical Imaging
- Computer Vision
- Pediatric Radiology
Background:
- Accurate skeletal growth estimation is crucial for pediatric health assessment.
- Manual assessment of skeletal growth from radiographs can be time-consuming and subjective.
Purpose of the Study:
- To develop and evaluate an automated knowledge-based vision system for skeletal growth estimation in children.
- To compare the system's accuracy with that of expert pediatric radiologists.
Main Methods:
- Utilized hand radiographs from 32 male and 25 female children (age 1-16 years).
- Employed hierarchical inferences and active shape models for automatic phalanx bone localization and segmentation.
- Extracted shape descriptors from bone contours to quantify skeletal growth.
- Applied a regression model and a Bayesian estimator using a selected feature vector.
Main Results:
- Achieved an estimation accuracy of 84% for females and 82% for males.
- Demonstrated accuracy comparable to expert pediatric radiologists.
- The automated system successfully quantified skeletal growth from bone contours.
Conclusions:
- The automated vision system shows significant potential for reliable skeletal growth estimation in children.
- This technology could aid pediatricians in clinical decision-making and monitoring growth disorders.
- The system offers a promising, objective alternative to manual radiograph assessment.