BİLDİRİLER

BİLDİRİ DETAY

Hıdır Selçuk NOĞAY
BREAST CANCER DIAGNOSIS BY USING DEEP LEARNING
 
Breast cancer is quite common and affects women worldwide. It is estimated that approximately 2.3 million new cases are diagnosed each year. However, thanks to early diagnosis and advances in medical science, survival rates have been increasing significantly over the years. In terms of severity, breast cancer can range from early stage, which is highly treatable, to advanced stage, which may require more intensive treatment. It is very important to undergo regular screenings and self-exams to detect possible abnormalities as early as possible. Diagnosing breast cancer typically involves a combination of techniques such as mammogram, ultrasound, biopsy, and genetic testing, depending on the specific case. These diagnostic tools help medical professionals determine the presence, stage, and characteristics of cancer. When it comes to treatment, there are various approaches available depending on the individual's condition and the stage of the cancer. Common treatments include surgery, radiation therapy, chemotherapy, targeted therapy, and hormone therapy. The treatment plan is usually tailored to the patient's specific needs, taking into account factors such as tumor size and location, general health status, and personal preferences. By using artificial intelligence techniques, early diagnosis of breast cancer may be possible and early treatment can be started. Using a deep learning technique trained and tested with ultrasound images could be incredibly valuable for diagnosing and classifying breast cancer. Ultrasound imaging plays an important role in the early detection and monitoring of breast cancer, and leveraging deep learning can improve accuracy and efficiency in this process. In fact, a deep learning model-supported diagnostic system could have a significant impact on patient outcomes, ensuring timely intervention and improving survival rates. Additionally, such studies could contribute to the creation of a comprehensive database of ultrasound images that could help train future AI models and support ongoing research. In this study, breast cancer diagnosis and classification was developed with a convolutional neural network model (CNN) trained and tested with ultrasound images. As a result of the study, it was revealed that more reliable and higher accuracy models can be designed by using CNN effectively and increasing its generalizability. ORCID NO: 0000-0001-9105-508X

Anahtar Kelimeler: CNN, Transfer Learning, Deep Learning, Breast Cancer



 


Keywords: