Obtaining a Well-Trained Artificial Intelligence Algorithm from Cross-Validation in Endoscopy - Publikacja - MOST Wiedzy

Wyszukiwarka

Obtaining a Well-Trained Artificial Intelligence Algorithm from Cross-Validation in Endoscopy

Abstrakt

The article shortly discusses endoscopic video analysis problems and artificial intelligence algorithms supporting it. The most common method of efficiency testing of these algorithms is to perform intensive cross-validation. This allows for accurately evaluate their performance of generalization. One of the main problems of this procedure is that there is no simple and universal way of obtaining a specific instance of a well-trained algorithm which has efficiency comparable to efficiency suggested by cross-validation. In this paper, a method resolving this problem (at some circumstances) is proposed and examined in the task of recognizing cancer, healthy tissue, blurred frames and sharp frames on endoscopic videos by two exemplary artificial intelligence algorithms designed for this task, using neural networks and support vector machines. The results show that proposed method allows to obtain algorithms trained a little better results than the average algorithm after cross--validation, without requiring any additional testing nor training.

Cytuj jako

Pełna treść

pełna treść publikacji nie jest dostępna w portalu

Słowa kluczowe

Informacje szczegółowe

Kategoria:
Publikacja monograficzna
Typ:
rozdział, artykuł w książce - dziele zbiorowym /podręczniku w języku o zasięgu międzynarodowym
Tytuł wydania:
W: System Analysis Approach to the Design, Control and Decision Support strony 159 - 168
Język:
angielski
Rok wydania:
2014
Opis bibliograficzny:
Cychnerski J., Brzeski A., Rościszewski P.: Obtaining a Well-Trained Artificial Intelligence Algorithm from Cross-Validation in Endoscopy// W: System Analysis Approach to the Design, Control and Decision Support/ ed. 978-83-7493-857-0 Wrocław: Oficyna Wydawnicza Politechniki Wrocławskiej, 2014, s.159-168
Weryfikacja:
Politechnika Gdańska

wyświetlono 28 razy

Publikacje, które mogą cię zainteresować

Meta Tagi