Published Paper Details:

Thyroid nodules, which are characterised as aberrant thyroid cell growth, might be an indicator of an excessive iodine intake, thyroid degeneration, inflammation, and other diseases. Despite the fact that thyroid nodules are mostly benign, the likelihood that they are malignant rises noticeably every year. The labour of medical practitioners is reduced and unnecessary fine needle aspiration and surgical excision are avoided and thanks to a number of studies that have been done to identify thyroid nodules using deep-learning-based image recognition analysis. A novel deep learning architecture is proposed in this study to reliably identify benign and malignant thyroid nodules. We imported this data from Algeria's hospitals exactly in Setifcity, and it was labelled by volunteer doctors. There are 1546 images belonging to three different classes namely normal thyroid, benign and malignant nodules where benign referring to less harmful and malignant being highly dangerous. First, a ResNet50 model pretrained on ImageNet database was applied to a collection of ultrasound thyroid images.The trained model has now able to classify the thyroid nodules into different categories like normal, benign and malignant. Overall, the proposed model demonstrated that ultrasound images and deep learning may be used to distinguish between benign and cancerous thyroid nodules.