Thyroid-Disease-Classification

Thyroid Disease Classification using Ultrasound Images

The thyroid gland in the human body is responsible for producing and regulating the metabolism of the body and hence are a very import gland in our body and thus requires high care and nurturing.The types of disorders that can occur because of the low or high secretion of the hormones by the thyroid gland which leads to various abnormalities and difference in shape and size. Another reason for the project was scale as there are around 42 million people in India suffering from different kind of abnormalities.The existing system which is put in place to detect and rectify diseases related to the thyroid gland are either absolute or use the image based analysis of the gland.The image based analysis of the thyroid gland and tissue is faulty in the way that it produces low resolution images and thus the subjective knowledge of the doctor is limited in analyzing the thyroid ultrasound. So in this project we have proposed a robust approach towards solving the above stated problem through the use of Machine Learning.The model proposed by us comprises of the image acquisition, Preprocessing the images(converting images to array) followed by segmentation of the images using K-means clustering and using the convolutional neural networks to train and test and model which would help in classification as well as the localisation of the disease part in the ultrasound image of the thyroid gland.The project was tested on a dataset of the ultrasound images as well as UCI machine learning dataset was used to further test the model and for future work related to thyroid ultrasound.