講演情報
[14p-P07-57]GLCM-based Region Boundary Identification in Shoulder Ultrasound
〇(M1)Jinyi Zhao1, Mizuki Fujiwara2,3, Masahiro Yamaguchi1, Marie Tabaru1 (1.Science Tokyo, 2.Tohoku Univ., 3.JR Sendai Hospital)
キーワード：
Ultrasound imaging、GLCM、Region segmentation
This study addresses frozen shoulder diagnosis using ultrasound imaging, moving beyond manual ROI placement by applying GLCM texture analysis to identify tissue boundaries. Texture features like homogeneity, energy, and mean gray level are extracted and visualized as pseudo-color maps, aiding observation and serving as foundational training data for future machine learning models.
By combining results from different directional analyses, the study segments ultrasound images into four regions: skin, deltoid muscle, fascia, and subscapularis & others. This segmentation introduces "thickness" as a diagnostic parameter, enhancing the ability to locate boundaries and assess tissue adhesion.
The findings establish a strong basis for automated segmentation and motion analysis, paving the way for more accurate and efficient diagnostic methods.
By combining results from different directional analyses, the study segments ultrasound images into four regions: skin, deltoid muscle, fascia, and subscapularis & others. This segmentation introduces "thickness" as a diagnostic parameter, enhancing the ability to locate boundaries and assess tissue adhesion.
The findings establish a strong basis for automated segmentation and motion analysis, paving the way for more accurate and efficient diagnostic methods.