Tanısal Ultrasonografi-Derin Öğrenme-Yapay Zek, Durur Subaşı I,Ateş HF, Editör, Türkiye Klinikleri Yayınevi, İstanbul, ss.66-72, 2022
ABS TRACT Artificial intelligence (AI) is a computer algorithm and can be defined as the ability
to perform processes such as reasoning, learning, adaptation, sensory understanding, and interaction,
which are related to human intelligence, and it is a whole of different subsystems. These subsystems
are machine learning, convolutional neural networks, and deep learning. Studies on the
use of artificial intelligence in radiology have increased significantly in the last decade. Breast radiology
is a prominent field in these artificial intelligence studies. Computer-assisted diagnosis
(CAD) found one of its first uses in breast radiology and has been in the use of breast radiology since
the 1990s which was initiated with simple algorithms. CAD is divided into three main groups as
CADe, CADx and CADt, as detection, decision support and triage respectively. Although AI research
and its commercial use first developed in the form of CAD in mammography, that is, for cancer
detection, several AI researches for breast ultrasound can be found in the literature. The clinical
use of CADe for lesion detection in automated breast ultrasound imaging, and as a diagnostic decision
support mechanism in handheld breast ultrasound is possible. In handheld breast ultrasound,
on the other hand, it is difficult to develop CADe systems because the scanning with the probe is
user dependent and cannot yet be standardized. However, in the future, standardized probe monitoring
will be possible and CADe algorithms will be able to be run in real time. On the other hand,
high false positivity is the main weakness of breast ultrasound and can be reduced with add-on effective
CADx algorithms. In the future, AI will enable the development of better image processing
algorithms, application of different