National and Kapodistrian University of Athens
Διερεύνηση αλλοιώσεων του στομάχου με τεχνική τεχνητής νοημοσύνης τύπου ακτινικής βάσης σε κυτταρολογικό υλικό
Abstract
dc:descriptionIntroduction A customized image analysis system has been developed and used to measure 25 parameters characterizing the shape, the size and the chromatine texture of nuclei from benign or malignant cells. Matterial and Methods In this study there have been studied benign and malignant cases of the stomachThe cytological examination of all the cases has been confirmed by the histological examination and by the clinical confirmation. For each organ we have attempted to discriminate between benign and malignant. There have been used τςο different types of neural networks, specifically the back propagation of the error (BP) neural network and the Radial Basis Functions (RBF) type neural network. Results The results indicate that it is possible to achieve excellent diagnostic accuracy. This fact confirms the proposed methodology, introduced in this study, in relation to the selected measurement characteristics and the selected artificial intelligence algorithms. These results are encouraging and indicate that our methodology could be useful in the everyday practice of the cytological laboratory. Despite the proposed method introduces and objectivity to the decision making process, due to the measurements, it is obvious that even in the sub-microscopic level there is an overlap in the characteristics of the feature space. This confirms the shortcoming of the human vision and intelligence to correctly discriminate the neoplastic nature of a lesion using only the cytological-morphological characteristics. Discussion In short the proposed method is an new “tool” in the scientific “quiver” of cytology, due to its contribution in the following: · Reduces the controversial cytological diagnoses · Improves the diagnostic accuracy of the cytological diagnosis. · Introduces an objectivity into the diagnostic methodology by the employment of morphometric and densitometric measurements. · Provides support to the decision making process of young cytologists that do not have enough experience and specialization at specific human organs. Finally it seems that the introduction of this methodology in the every day routine of the cytological laboratory could improve the pre-operational control of patients and could eventually reduce the not necessary surgeries. The “employment” of digital material and modern computers is ideal for the usage of this method both in real time and in second time (off-line), for the provisioning of the primary diagnosis or for the provisioning of consultation both locally and from distance.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2009
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tzivras, Dimitrios
- Τζιβράς, Δημήτριος
Subjects
dc:subject × 18- Αλλοιώσεις καλοήθεις κακοήθεις από το στόμαχο
- Κυτταρολογική εξέταση
- Ιστολογική εξέταση
- Τεχνητή νοημοσύνη
- Νευρωνικά δίκτυα
- Νευρωνικό δίκτυο οπισθοδρομική διάδοση σφάλματος
- Δίκτυο ακτινικών βασικών συναρτήσεων
- Benign malignant lesions stomach
- Cytology
- Histology
- Artificial intelligence
- Neural networks
- Back propagation of the error - BP
- Radial basis function - RBF
- Ιατρική και Επιστήμες Υγείας
- Βασική Ιατρική
- Medical and Health Sciences
- Basic Medicine
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/23803