DocumentCode :
724954
Title :
Automated scoring of fetal abdomen ultrasound scan-planes for biometry
Author :
Chandan Kumar, A.M. ; Shriram, K.S.
Author_Institution :
GE Global Res., Bangalore, India
fYear :
2015
fDate :
16-19 April 2015
Firstpage :
862
Lastpage :
865
Abstract :
Biometry on the Ultrasound images of the fetal abdomen aids in determining fetal well being. Finding standard scanplane requires the human operator to possess a high degree of skill/expertise. This paper describes an automatic method for scan-plane identification of the fetal abdomen based on the spatial configuration of key anatomical landmarks - spine, stomach and vein. A novel Biometry Suitability Index (BSI) using the detected anatomical landmarks indicates whether a scan-plane can be used for biometry. A training database of 90 images was used to build a spatial configuration model the of the landmarks. Employing relatively simple individual detectors and relying primarily on the configuration model, the landmarks were located and tracked across frames/scanplanes in videos. The Biometry Suitability Index computed over video sequences was seen to closely mimic the expert´s assessment of scan-plane quality for biometry.
Keywords :
biomedical ultrasonics; blood vessels; image sequences; medical image processing; object detection; obstetrics; video signal processing; BSI; Biometry Suitability Index; anatomical landmarks; automated scoring; automatic method; fetal abdomen ultrasound scan-planes; fetal well being; human operator; scan-plane identification; scan-plane quality; spatial configuration model; spine; standard scan-plane; stomach; training database; ultrasound images; vein; video sequences; Abdomen; Indexes; Stomach; Ultrasonic imaging; Veins; Video sequences; Videos; Biometry; Fetal Abdomen; Scanplane; Ultrasound;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Biomedical Imaging (ISBI), 2015 IEEE 12th International Symposium on
Conference_Location :
New York, NY
Type :
conf
DOI :
10.1109/ISBI.2015.7164007
Filename :
7164007
Link To Document :
بازگشت