Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Research › peer-review
On the analysis of thyroid gland ultrasonic images by the methods of Haralic and local density function. / Ампилова, Наталья Борисовна; Соловьев, Игорь Павлович; Лямин, Владимир Андреевич.
On the analysis of thyroid gland ultrasonic images by the methods of Haralic and local density function. София : Софийский технический университет, 2025. p. 11-15.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Research › peer-review
}
TY - GEN
T1 - On the analysis of thyroid gland ultrasonic images by the methods of Haralic and local density function
AU - Ампилова, Наталья Борисовна
AU - Соловьев, Игорь Павлович
AU - Лямин, Владимир Андреевич
N1 - Conference code: 19
PY - 2025/10/21
Y1 - 2025/10/21
N2 - Ultrasonic diagnostic is the simple and effective method for detection of anomalies of thyroid gland, and in particular nodules.— new growths which are different from health tissue by. eсhogenity .There are 3 types of nodules iso-,hyper- and inechogenic. Isoehogenic nodules have the same density as the health tissue, and may be poorly distinguishable in the image. Two other types differ from health tissue by light.To reveal nodules we apply the machine learning method based on using Haralick texture features and multifractal spectrum. The image is represented by a set of non-overlapping blocks containing both normal tissue and nodules. The blocks are determined after preliminary markup performed by a medical specialist. Texture features are caculated for each block. For machine learning SVM method was taken as a model, Thanks to the block-based approach, the application implements the capability to localize a nodule within the image.
AB - Ultrasonic diagnostic is the simple and effective method for detection of anomalies of thyroid gland, and in particular nodules.— new growths which are different from health tissue by. eсhogenity .There are 3 types of nodules iso-,hyper- and inechogenic. Isoehogenic nodules have the same density as the health tissue, and may be poorly distinguishable in the image. Two other types differ from health tissue by light.To reveal nodules we apply the machine learning method based on using Haralick texture features and multifractal spectrum. The image is represented by a set of non-overlapping blocks containing both normal tissue and nodules. The blocks are determined after preliminary markup performed by a medical specialist. Texture features are caculated for each block. For machine learning SVM method was taken as a model, Thanks to the block-based approach, the application implements the capability to localize a nodule within the image.
M3 - Conference contribution
SN - 1314-2100
SP - 11
EP - 15
BT - On the analysis of thyroid gland ultrasonic images by the methods of Haralic and local density function
PB - Софийский технический университет
CY - София
Y2 - 20 October 2025 through 21 October 2025
ER -
ID: 142930170