Enhancing Maternofetal Ultrasound Images Toward Boosting Classification Performance on a Diverse and Comprehensive Data

Accurately classifying maternofetal ultrasound images is crucial for diagnosing fetal abnormalities and ensuring the mother and fetus’s health. This study proposes a novel approach for maternofetal ultrasound image classification using a Convolutional Neural Network (CNN) architecture and image preprocessing techniques. We utilized a pre-trained deep neural network on a large image dataset and applied image preprocessing techniques, such as the CLAHE and Unsharp filters, to enhance the quality of the images. Our findings show that the model achieved an impressive overall accuracy of 99.13% on the test set, demonstrating its effectiveness in accurately classifying maternofetal ultrasound images into their respective standard plans. Moreover, the CLAHE and Unsharp filters significantly improved the accuracy of the model by enhancing the image quality. This study highlights the potential of deep learning models in improving medical imaging diagnosis and offers a promising approach for accurate and efficient maternofetal ultrasound image classification.