Multimodal Feature Fusion and Knowledge-Driven Learning via Experts Consult for Thyroid Nodule Classification

Computer-aided diagnosis (CAD) is becoming a prominent approach to assist clinicians spanning across multiple fields. These automated systems take advantage of various computer vision (CV) procedures, as well artificial intelligence (AI) techniques, formulate given image, e.g., computed tomography and ultrasound. Advances in both areas (CV AI) are enabling ever increasing performances CAD systems, which can ultimately avoid performing invasive procedures such fine-needle aspiration. In this study, novel end-to-end knowledge-driven classification framework presented. The system focuses on multimodal data generated by thyroid ultrasonography, acts providing nodule into the benign malignant categories. Specifically, proposed leverages cues provided an ensemble experts guide learning phase densely connected convolutional network (DenseNet). composed networks pretrained ImageNet, including AlexNet, ResNet, VGG, others. previously feature parameters used create ultrasonography domain via transfer learning, decreasing, moreover, number samples required for training. To validate method, extensive experiments were performed, detailed DenseNet. As demonstrated results, achieves relevant terms qualitative metrics task, thus resulting great asset when formulating diagnosis.