Paper Info
Title
Surface-based registration of liver in ultrasound and CT
Abstract
Ultrasound imaging is an attractive modality for real-time image-guided interventions. Fusion of US imaging with a diagnostic imaging modality such as CT shows great potential in minimally invasive applications such as liver biopsy and ablation. However, significantly different representation of liver in US and CT turns this image fusion into a challenging task, in particular if some of the CT scans may be obtained without contrast agents. The liver surface, including the diaphragm immediately adjacent to it, typically appears as a hyper-echoic region in the ultrasound image if the proper imaging window and depth setting are used. The liver surface is also well visualized in both contrast and non-contrast CT scans, thus making the diaphragm or liver surface one of the few attractive common features for registration of US and non-contrast CT. We propose a fusion method based on point-to-volume registration of liver surface segmented in CT to a processed electromagnetically (EM) tracked US volume. In this approach, first, the US image is pre-processed in order to enhance the liver surface features. In addition, non-imaging information from the EM-tracking system is used to initialize and constrain the registration process. We tested our algorithm in comparison with a manually corrected vessel-based registration method using 8 pairs of tracked US and contrast CT volumes. The registration method was able to achieve an average deviation of 12.8mm from the ground truth measured as the root mean square Euclidean distance for control points distributed throughout the US volume. Our results show that if the US image acquisition is optimized for imaging of the diaphragm, high registration success rates are achievable.
Year
DOI
Venue
2015
10.1117/12.2082160
Proceedings of SPIE
Keywords
Field
DocType
Multi-modality fusion,image registration,liver-diaphragm surface,image-guided intervention,ultrasound,computed tomography,electromagnetic tracking
Computer vision,Diaphragm (structural system),Image fusion,Liver biopsy,Medical imaging,Euclidean distance,Ground truth,Artificial intelligence,Image registration,Ultrasound,Physics
Conference
Volume
ISSN
Citations 
9415
0277-786X
0
PageRank 
References 
Authors
0.34
3
9
Name
Order
Citations
PageRank
Ehsan Dehghan18310.60
Kongkuo Lu2809.23
Pingkun Yan3130683.14
Amir M. Tahmasebi4609.66
Sheng Xu550771.47
Bradford J Wood614231.69
Nadine Abi-Jaoudeh7143.02
aradhana m venkatesan822.12
Jochen Kruecker916115.19