Abstract | ||
---|---|---|
3D echocardiography has enabled new clinical applications of ultrasound related to both interventional guidance and quantification of chamber characteristics (e.g. volumes, function). However, image quality may be hampered by dropouts in the image as well as limited field of view. By compounding data from several overlapping images, a volume with extended field of view can be formed. A 3D method based on Farneback optic flow is proposed to perform registration between ultrasound images taken from different orientations. It utilizes signal decomposition into polynomial basis functions and solves the transformation between the volumes analytically. Validation using synthetic data sets showed a registration error of 0.47 +/- 0.05 mm. And testing on data sets of 50 real images showed promising results. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1007/978-3-319-20309-6_15 | Lecture Notes in Computer Science |
Keywords | Field | DocType |
Rigid registration of 3D echocardiography,Extension of field of view,Farneback optic flow | Field of view,Computer vision,Data set,Computer science,Flow (psychology),Polynomial basis functions,Image quality,Artificial intelligence,Real image,Synthetic data sets,Ultrasound | Conference |
Volume | ISSN | Citations |
9126 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 9 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
A. Danudibroto | 1 | 0 | 0.34 |
Olivier Gérard | 2 | 0 | 0.34 |
Martino Alessandrini | 3 | 165 | 16.19 |
Oana Mirea | 4 | 10 | 1.73 |
Jan D'hooge | 5 | 42 | 4.73 |
Eigil Samset | 6 | 133 | 16.57 |