Browse Prior Art Database

Real-time 3D Reconstruction for Collision Avoidance in Interventional Environments

IP.com Disclosure Number: IPCOM000173410D
Published in the IP.com Journal: Volume 8 Issue 8B (2008-08-25)
Included in the Prior Art Database: 2008-Aug-25
Document File: 2 page(s) / 60K

Publishing Venue

Siemens

Related People

Juergen Carstens: CONTACT

Abstract

Nowadays, one or more automated devices, such as C-arms and medical robots, are used in an interventional room. However, these devices and the other equipments such as ceiling-suspended monitor arms or radiation protection shields can collide. A collision can damage both the device and the colliding object so that both have to be taken out of service until a technician has evaluated the damage. Further, the intervention has to be suspended and the patient may be moved to another room. Thus, such collisions can be dangerous and costly. Up to now, a fully automated system for collision avoidance is not available. Most systems rely on the discretion of the operating physician to avoid collisions. The technical measurements comprise of contact sensors and reduced movement speed.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 43% of the total text.

Page 1 of 2

Real-time 3D Reconstruction for Collision Avoidance in Interventional Environments

Idea: Dr. Thomas Redel, DE-Forchheim; Alexander Ladikos, DE-München; Dr. Selim Ben Himane,
DE-München; Prof. Dr. Nassir Navab, DE-München

Nowadays, one or more automated devices, such as C-arms and medical robots, are used in an interventional room. However, these devices and the other equipments such as ceiling-suspended monitor arms or radiation protection shields can collide. A collision can damage both the device and the colliding object so that both have to be taken out of service until a technician has evaluated the damage. Further, the intervention has to be suspended and the patient may be moved to another room. Thus, such collisions can be dangerous and costly. Up to now, a fully automated system for collision avoidance is not available. Most systems rely on the discretion of the operating physician to avoid collisions. The technical measurements comprise of contact sensors and reduced movement speed.

Therefore, a novel real-time collision avoidance system is proposed which increases the safety in the interventional room and allows faster device operation. The proposed system consists of 16 optical cameras mounted on the ceiling in the interventional room and Personal Computers (PC) performing the reconstruction and collision avoidance. In the offline phase the working volume of the device is determined, the cameras are installed and some background images are acquired. In the online phase, the background images are used to segment the foreground objects. The segmented images are subsequently used to perform the 3D reconstruction of the visual hull of the objects in the interventional room. In a final step, safety zones are computed around the objects and checked if they intersect with the movement range of the device. A more detailed version of the proposed reconstruction system is described in the following:

• System Architecture: The cameras are mounted on the ceiling. To ensure that all cameras take images at the same time, the cameras are externally triggered. Four workstations and one master PC are used to perform the reconstruction. The PCs are placed in a neighboring room so that in the interventional room additional space is not occupied. The data flow of the system is depicted in Figure 1.

• Calibration: Before the system can be used, the cameras are calibrated and registered according to the room coordinate system. For this purpose a method is used which relies on point correspondence created by a point light source such as a LED. First, the light of the room is dimmed so that the point is easy to detect. Then the light source has to be moved through the reconstruction volume in order to create the correspondences used in the calibration. Finally, the camera coordinate system is registered to the room co...