A state-of-the-art new method for diagnosing atrial septal defects with origami technique augmented dataset and a column-based statistical feature extractor

• Published in: Computers in biology and medicine Vol. 196; no. Pt C; p. 110967
• Main Authors: Yaman, Irfan, Kilic, Irfan, Yaman, Orhan, Poyraz, Fatih, Erdem Kaya, Emin, Ozgur Baris, Veysel, Ciris, Sukru
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.09.2025
• Subjects: Adult; Atrial septal defect (ASD); Chest X-ray (CXR); Column-based statistics extractor; Databases, Factual; Female; Heart Septal Defects, Atrial - diagnostic imaging; Humans; Image classification; Male; Origami data augmentation; Support Vector Machine; Support vector machine (SVM); Support vector machine (SVM); Atrial septal defect (ASD); Column-based statistics extractor; Chest X-ray (CXR); Origami data augmentation; Image classification
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2025.110967

Early diagnosis of atrial septal defects (ASDs) from chest X-ray (CXR) images with high accuracy is vital. This study created a dataset from chest X-ray images obtained from different adult subjects. To diagnose atrial septal defects with very high accuracy, which we call state-of-the-art technology, the method known as the Origami paper folding technique, which was used for the first time in the literature on our dataset, was used for data augmentation. Two different augmented data sets were obtained using the Origami technique. The mean, standard deviation, median, variance, and skewness statistical values were obtained column-wise on the images in these data sets. These features were classified with a Support vector machine (SVM). The results obtained using the support vector machine were evaluated according to the k-nearest neighbors (k-NN) and decision tree classifiers for comparison. The results obtained from the classification of the data sets augmented with the Origami technique with the support vector machine (SVM) are state-of-the-art (99.69 %). Our study has provided a clear superiority over deep learning-based artificial intelligence methods. •Data augmentation technique using origami art is demonstrated.•The column-based statistical feature extraction method is presented.•Chest X-ray (CXR) datasets for atrial septal defect (ASD) diagnosis have been added to the literature.•A new generation method superior to deep learning methods is proposed.