Artificial Intelligence-Based Thyroid Nodule Classification Using Information from Spatial and Frequency Domains

Abstract

Image-based computer-aided diagnosis (CAD) systems have been developed to assist doctors in the diagnosis of thyroid cancer using ultrasound thyroid images. However, the performance of these systems is strongly dependent on the selection of detection and classification methods. Although there are previous researches on this topic, there is still room for enhancement of the classification accuracy of the existing methods. To address this issue, we propose an artificial intelligence-based method for enhancing the performance of the thyroid nodule classification system. Thus, we extract image features from ultrasound thyroid images in two domains: spatial domain based on deep learning, and frequency domain based on Fast Fourier transform (FFT). Using the extracted features, we perform a cascade classifier scheme for classifying the input thyroid images into either benign (negative) or malign (positive) cases. Through expensive experiments using a public dataset, the thyroid digital image database (TDID) dataset, we show that our proposed method outperforms the state-of-the-art methods and produces up-to-date classification results for the thyroid nodule classification problem.

Ultrasound Image-Based Diagnosis of Malignant Thyroid Nodule Using Artificial Intelligence

“…The malign case can cause thyroid cancer. Fortunately, most detected thyroid nodules are benign cases as reported in the previous studies [19,36]. However, diagnosing and treating malign cases is still very important.…”

“…They could be caused by many factors, including iodine deficiency, overgrowth of normal thyroid tissue, or thyroid cancer. Thyroid nodules are usually classified into two categories based on their characteristics namely, benign cases (which are noncancerous nodules), and malign cases (which can cause thyroid cancer) [36]. In both the benign and malign cases, the appearance of thyroid nodules can cause problems with patient health.…”

“…Using the ultrasound thyroid nodule images, there have been several previous studies on CAD for the thyroid nodule detection and classification problems. In contrast to the conventional thyroid diagnosis methods mentioned above, the CAD methods for thyroid nodules use ultrasound thyroid nodule images as inputs and produce thyroid nodule regions and/or the status of nodules (benign or malign) [17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36]. Similar to normal image processing systems, CAD systems for thyroid nodule use several image processing techniques to extract information from input images for detection/classification purposes.…”

“…With the development of technology, such as the back propagation algorithm, neural network, and Graphics Processing Units (GPUs), the deep learning-based technique has recently been applied to solve many problems in medical image processing systems [1,2,10,12,21,36]. For the thyroid nodule detection/classification problem, the deep learning-based method has gained a lot of success.…”

“…As a fusion of the two mentioned approaches, there exist studies that combine the handcrafted and deep learning methods to enhance the classification performance. In a study by Nguyen et al [36], they found that the information in the frequency domain can be useful for discriminating easy samples of benign and malign cases, and the deep learning-based method can be useful for discriminating harder samples (ambiguous samples). Based on this observation, they proposed a method that applies a cascade classifier scheme for the thyroid nodule classification problem.…”