Principles of creating a decision-making system for a radiologist-mammologist
Вантажиться...
Дата
2021
Автори
Назва журналу
Номер ISSN
Назва тому
Видавець
Анотація
For the main method of breast cancer prevention-screening, there is
currently no CAD system that can work on the "white box" principle. False positives
also reduce the confidence of doctors with the CAD system. Also, there are no systems
for monitoring the laying of the breast for an X-ray technician. The development of
CAD systems in mammology is mainly carried out using computer vision technologies,
while the proposed solution uses not only computer vision but also a knowledge base
for verifying the output of neural networks and production output, taking into account
the probabilities of the appearance of signs of the disease. This approach is quite
flexible and scalable – knowledge can be both accumulated and obtained, and
recognition modules can be added or retrained. Changing the knowledge base allows
you to show the doctor based on which knowledge the pathological formation was
determined. But most importantly, knowledge base agents in the future can connect to
the hospital information system (HIS) and independently request the data necessary
for the preparation of the study report. This approach allows you to fully simulate the
activities of a diagnostic doctor.
Опис
Efimova A. Principles of creating a decision-making system for a radiologist-mammologist / A. Efimova ; advisor : A. Kayeshko // Black Sea Science 2021 : proc. of the Intern. Competition of Student Scientific Works / Odessa National Academy of Food Technologies ; eds. B. Yegorov, M. Mardar [et al.]. – Odessa: ONAFT, 2021. – P. 383–390 : fig. – Ref.: 8 tit.
Ключові слова
breast, cancer, CAD system, knowledge base, radiologist-mammologist, OSTIS, neuron networks, white box, nosological form, agents