No abstract

Abstract. Automatic and robust registration of pre-operative magnetic resonance imaging (MRI) and intra-operative ultrasound (US) is essential to neurosurgery. We reformulate and extend an approach which uses a Linear Correlation of Linear Combination (LC 2 )-based similarity metric, yielding a novel algorithm which allows for fully automatic US-MRI registration in the matter of seconds. It is invariant with respect to the unknown and locally varying relationship between US image intensities and both MRI intensity and its gradient. The overall method based on this both recovers global rigid alignment, as well as the parameters of a free-form-deformation (FFD) model. The algorithm is evaluated on 14 clinical neurosurgical cases with tumors, with an average landmark-based error of 2.52 mm for the rigid transformation. In addition, we systematically study the accuracy, precision, and capture range of the algorithm, as well as its sensitivity to different choices of parameters.

In this paper we present two efficient GPU-based visual hull computation algorithms. We compare them in terms of performance using image sets of varying size and different voxel resolutions. In addition, we present a real-time 3D reconstruction system which uses the proposed GPUbased reconstruction method to achieve real-time performance (30 fps) using 16 cameras and 4 PCs.

Abstract. With the increased presence of automated devices such as C-arms and medical robots and the introduction of a multitude of surgical tools, navigation systems and patient monitoring devices, collision avoidance has become an issue of practical value in interventional environments. In this paper, we present a real-time 3D reconstruction system for interventional environments which aims at predicting collisions by building a 3D representation of all the objects in the room. The 3D reconstruction is used to determine whether other objects are in the working volume of the device and to alert the medical staff before a collision occurs. In the case of C-arms, this allows faster rotational and angular movement which could for instance be used in 3D angiography to obtain a better reconstruction of contrasted vessels. The system also prevents staff to unknowingly enter the working volume of a device. This is of relevance in complex environments with many devices. The recovered 3D representation also opens the path to many new applications utilizing this data such as workflow analysis, 3D video generation or interventional room planning. To validate our claims, we performed several experiments with a real C-arm that show the validity of the approach. This system is currently being transferred to an interventional room in our university hospital.

In this paper we present an efficient and robust real-time system for object contour tracking in image sequences. The developed application partly relies on an optimized implementation of a state-of-the-art curve fitting algorithm, and integrates important additional features in order to achieve robustness while keeping the speed of the main estimation algorithm. An application program has been developed, which requires only a few standard libraries available on most platforms, and runs at video frame rate on a common PC with standard hardware equipment. Motivation and Scope of the Present WorkThe general problem of object contour tracking in image sequences is an important and challenging topic in the computer vision community; as many researchers already pointed out, an advanced contour tracking technique can provide crucial information for many image understanding problems and, at the same time, allows the development of efficient and useful working applications in many fields of interest. We refer the reader to [1] for a survey of these applications and the related references.Among the currently available methodologies, a very appealing one is the Contracting Curve Density (CCD) algorithm: this method has been recently developed and presented in [3] as a state-of-the-art improvement over other advanced techniques such as the Condensation algorithm [6], and it has been shown to overperform them in many different estimation tasks. Nevertheless, its higher complexity has been initially considered as an obstacle to an effective real-time implementation, even in the simplified form named Real-Time CCD from the same authors [4], and a working online version still had to be investigated.Moreover, in order to realize an autonomous and robust tracking system, some important additional issues have to be taken into account; one of them is the possibility of automatic initialization and reinitialization of the system, both at the beginning of the tracking, and in case of tracking loss. Nevertheless, one also expects that a "well behaving" tracking system does not need a global estimation procedure during most of the tracking task, or, in the ideal case, only at the beginning. The global initialization module is computationally more demanding than the online tracker, and a frequent reinitialization would significantly slow down the performance of the system. Therefore, most of the efforts in realizing such a system have to be focused on a robust real-time performance of the contour tracker itself, in terms of speed, accuracy and convergence area for the parameter search.All of the above mentioned issues, and other relevant aspects, constitute the main motivation of the present work; it will be shown how it contributes to the state-of-the-art in contour tracking systems with the following achievements:• The real-time CCD algorithm [4] has been reimplemented with a significant number of critical speedups with respect to the originally proposed version.• A global initialization and reinitialization module has been developed ...