Scheduled System Maintenance:
On May 6th, system maintenance will take place from 8:00 AM - 12:00 PM ET (12:00 - 16:00 UTC). During this time, there may be intermittent impact on performance. We apologize for the inconvenience.
By Topic

6-DOF pose estimation from single Ultrasound image using 3D IP models

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

The purchase and pricing options are temporarily unavailable. Please try again later.
5 Author(s)
Bo Zheng ; Ikeuchi Lab., Tokyo Univ., Tokyo ; Ishikawa, R. ; Oishi, T. ; Takamatsu, J.
more authors

Real-time pose estimation of a free-hand ultrasound (US) image without any position sensor is desirable for diagnostics and image guidance, but it suffers from poor image quality as a result of processing. The paper confronts this problem by proposing a new 6-DOF pose estimation method based on a fast registration process making use of 3D implicit polynomials (IP). The proposed registration method has some main advantages over the traditional methods. First, our formulation is based on minimization of energy functional derived from IP gradient flow, and thus it is more efficient than traditional registration because it eliminates the cost for calculating point-wise correspondences. Second, the efficient calculation benefits from the property of IP having few coefficients, which means that both the gradient field and its transformation can be calculated in an extremely light manner. Third, applying a real time US image pose estimation, we demonstrate the capabilities of overcoming the limitations of unconstrained free hand US data, resulting in robust and fast pose estimation.

Published in:

Computer Vision and Pattern Recognition Workshops, 2008. CVPRW '08. IEEE Computer Society Conference on

Date of Conference:

23-28 June 2008