Statistical analysis of texture features of microcalcifications at mammographic images prior and post biopsy

Postgraduate Thesis uoadl:2882368 368 Read counter

Unit:
Κατεύθυνση Πληροφορική στην Ιατρική
Πληροφορική
Deposit date:
2019-10-09
Year:
2019
Author:
Duni Ergesta
Supervisors info:
Ιωάννης Καλατζής, αναπληρωτής καθηγητής, τμήμα μηχανικών Βιοιατρικής, Πανεπιστήμιο Δυτικής Αττικής (ΠΑΔΑ)
Original Title:
Στατιστική ανάλυση χαρακτηριστικών υφής σε μικροαποτιτανώσεις σε εικόνες μαστογραφίας πριν και μετά την διενέργεια βιοψίας
Languages:
English
Greek
Translated title:
Statistical analysis of texture features of microcalcifications at mammographic images prior and post biopsy
Summary:
The presence of microcalcifiacations in mammographies is considered to be the main early finding in most malignant cases, however microcalcifiacations are present in benign cases as well. This fact, sometimes, makes the distinction between two cases particularly challenging. In these cases the final diagnosis is made after biopsy. The aim of present study was to evaluate the texture features prior and post biopsy.
Texture features which were calculated are features derived from first-order and second-order statistics. The first-order features were mean value, standard deviation, skewness and kyrtosis. The second-order features, which were calculated, were contrast, correlation, entropy, and homogeneity.
Texture features were calculated on the ROIs of 47 patients prior and post biopsy. Statistical analysis was performed with Mann Whitney U test. The Octave program was used to calculate these features and in order to perform statistical analysis.
From the results obtained, it was found that before biopsy only the feauture of entropy differs significantly between two groups. After biopsy, a statistically significant difference is observed at more features. These features are entropy, standard deviation, homogeneity and contrast.
In conclusion, entropy appeared to be a feature that differs significantly both at mammographies, as well as at post-biopsy images. It was, also, found that more features differ significantly at post-biopsy images than at images prior the biopsy.
Main subject category:
Technology - Computer science
Keywords:
texture features, statistical analysis, microcalcifications, co-occurrence matrix, histogram
Index:
Yes
Number of index pages:
6
Contains images:
Yes
Number of references:
31
Number of pages:
74
File:
File access is restricted only to the intranet of UoA.

Duni.Ergesta.pdf
2 MB
File access is restricted only to the intranet of UoA.