DETECTION AND CLASSIFICATION OF BENIGN AND MALIGNANT THYROID USING CONVOLUTIONAL NEURAL NETWORK

Abstract

For the detection of thyroid, ultrasonography is used to imaging the thyroid nodules. Using
ultrasonographic images the classification of benign and malignant thyroid is inappropriate.
So convolutional neural networklayer is used for diagnosis of thyroid. Ultrasonogrphy can
detect only the presence and absence of thyroid nodules. So CNN algorithm is used to classify
whether the thyroid is at high or low level. The beginning stage of thyroid is benign stage and
the later stage is malignant one. The number datasets are collected and trained with CNN to
obtain certain accuracy level. The combination of convolution ultrasound and ultrasound
elasticity images is done based on convolutional neural network for obtaining a clear and
accuracy classification and segmentation. Here the accuracy level of 96.65% is obtained