Feature Extraction of Ultrasound Thyroid Images for Thyroid Cancer Detection

Thyroid-related disorders have emerged as a notable public health challenge on a global scale, emphasizing the need for advanced diagnostic techniques that offer precision and efficiency in disease detection. This research paper presents a comprehensive investigation focusing on feature extraction techniques applied to ultrasound (US) thyroid images, utilizing the MATLAB programming environment. The primary objective of this study is to elevate the accuracy and effectiveness of thyroid cancer detection, ultimately contributing to early diagnosis and enhancing patient care. The proposed research entails a series of image preprocessing steps performed in MATLAB, employing various preprocessing techniques. Following the preprocessing phase, the process of feature extraction involves the use of two primary methods: Histogram-based techniques and Gray-Level Co-occurrence Matrix. These extracted features will be leveraged in the subsequent phase to detect and classify thyroid cancer, employing machine learning classification techniques. This research serves as an important step in the ongoing efforts to develop more precise and efficient diagnostic tools for detecting thyroid cancer, ultimately leading to improved healthcare outcomes for patients.