Related Experiment Video
Updated: Sep 7, 2025
08:20
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
Published on: October 27, 2023
1.6K
An improved deep convolutional neural network architecture for chromosome abnormality detection using hybrid
N Nimitha1, P Ezhumalai2, Arun Chokkalingam1
1Department of ECE, RMK College of Engineering and Technology, Puduvoyal, India.
Microscopy Research and Technique
|June 16, 2022
Summary
Automated chromosome karyotyping using a deep convolutional neural network (DCNN) with generative adversarial networks and hybrid moth-flame optimization significantly improves accuracy and reduces time. This novel approach aids cytogenetic experts in diagnosing numerical chromosome abnormalities more efficiently.
Area of Science:
- Genetics and Genomics
- Computational Biology
- Medical Imaging Analysis
Background:
- Manual karyotyping is a complex, time-consuming, and error-prone method for detecting chromosome abnormalities.
- Existing automated methods often require large datasets and extensive hyperparameter tuning.
Purpose of the Study:
- To develop an automated chromosome karyotype architecture using deep convolutional neural networks (DCNNs).
- To address the challenges of limited datasets and complex hyperparameter tuning in automated karyotyping.
Main Methods:
- A DCNN architecture was employed, enhanced with generative adversarial networks (GANs) to create synthetic training data.
- Hyperparameter tuning was optimized using a hybrid moth-flame optimization integrated with a hill-climbing strategy (HMFOHC).
- The model was trained and validated on the BioImLab chromosome and hospital datasets.
Main Results:
- The proposed HMFOHC-optimized DCNN achieved high accuracy (98.65%), F1-score (98.86%), and kappa coefficient (0.9894) in multiclass classification of numerical chromosome abnormalities.
- The model successfully differentiated five numerical abnormalities: Trisomy 13, Trisomy 18, Trisomy 21, Trisomy XXY syndrome, and Monosomy X.
- Inference time was reduced to 12.5 seconds, outperforming existing state-of-the-art techniques.
Conclusions:
- The automated karyotyping system offers a significant improvement over manual methods in terms of speed and accuracy.
- This DCNN-based approach provides a robust and efficient tool for cytogenetic analysis and diagnosis.
- The methodology has the potential to assist forensic experts in making faster and more informed decisions.
