Breast cancer is the most common cancer worldwide. Early detection is crucial for reducing its mortality rate. Mammography is beneficial for early breast cancer diagnosis and reduces cancer mortality. Mammography is a useful tool for detecting breast cancer at an early stage, which can help reduce cancer-related deaths [1, 2]. Mammography guidelines define the necessary equipment, quality control, and techniques for proper imaging [3, 4]. Mammographic positioning is an essential factor that influences the extraction of lesions in imaging techniques, and no matter how high-performance imaging equipment is available, improper positioning cannot be supplemented. Mammography is a challenging procedure when it comes to positioning the breasts correctly. This is because the evaluation process is subjective and relies heavily on visual inspection. Visual and qualitative assessments for positioning are used to determine mammographic propriety; however, there is significant variation in visual assessment between individuals, leading to accuracy issues.
In recent years, positioning evaluation for mammography has been performed using artificial intelligence. However, the mammogram parts, such as the inframammary fold and nipple, are limited [5]. In this study, we propose quantitively evaluating mammographic retromammary space using a deep convolutional neural network in which mammograms can be detected automatically.
