Related Experiment Video
Updated: Jun 1, 2025
15:18
Making MR Imaging Child's Play - Pediatric Neuroimaging Protocol, Guidelines and Procedure
Published on: July 30, 2009
18.1K
Predicting Paediatric Brain Disorders from MRI Images Using Advanced Deep Learning Techniques.
Yogesh Kumar1, Priya Bhardwaj2, Supriya Shrivastav3
1Department of CSE, School of Technology, Pandit Deendayal Energy University, Gandhinagar, Gujarat, India.
Neuroinformatics
|January 17, 2025
Summary
This study introduces an AI system using deep learning models to accurately detect childhood diseases from MRI scans. The InceptionResNetV2 model achieved 97.59% accuracy, improving pediatric diagnosis.
Area of Science:
- Medical Imaging Analysis
- Artificial Intelligence in Healthcare
- Pediatric Diagnostics
Background:
- Children's diseases pose a global health challenge requiring timely and accurate diagnosis.
- Conventional diagnostic methods are often tedious, inaccurate, and lead to treatment delays.
- Artificial intelligence (AI), particularly deep learning, offers potential for improved medical image analysis.
Purpose of the Study:
- To develop and evaluate an AI-driven system for detecting childhood diseases using advanced Convolutional Neural Network (CNN) models.
- To compare the performance of various CNN architectures on pediatric brain disorder MRI images.
- To identify the most effective AI models for accurate and efficient disease diagnosis in children.
Main Methods:
- Utilized a range of CNN models including EfficientNetB0, Xception, InceptionV3, MobileNetV2, VGG19, DenseNet169, ResNet50V2, ResNet152V2, and InceptionResNetV2.
- Trained models on MRI images of pediatric brain disorders, employing data visualization techniques like segmentation and contour-based feature extraction.
- Optimized model performance using ADAM and RMSprop optimizers, evaluating metrics such as accuracy, loss, RMSE, precision, recall, and F1 score.
Main Results:
- The InceptionResNetV2 model, optimized with ADAM, achieved the highest accuracy of 97.59%.
- EfficientNetB0, optimized with RMSprop, attained 94.59% accuracy, while also demonstrating the lowest loss (0.25) and RMSE (0.5) when optimized with ADAM.
- Evaluations included precision, learning curves, recall, computational time, and F1 score, underscoring the efficacy of the AI approaches.
Conclusions:
- AI-driven deep learning models significantly enhance the accuracy and efficiency of diagnosing childhood diseases from medical images.
- The proposed system demonstrates the potential of advanced CNN architectures for improving pediatric healthcare outcomes.
- This AI approach offers a promising solution to overcome limitations of conventional diagnostic methods, enabling faster and more reliable disease detection.
