Elsevier

Medical Image Analysis

Volume 6, Issue 3, September 2002, Pages 289-300
Medical Image Analysis

Algorithmic tools for real-time microsurgery simulation,

https://doi.org/10.1016/S1361-8415(02)00086-5Get rights and content

Abstract

Today, there is growing interest in computer surgical simulation to enhance surgeons’ training. This paper presents a simulation system based on novel algorithms for animating instruments interacting with deformable tissue in real-time. The focus is on computing the deformation of a tissue subject to external forces, and detecting collisions among deformable and rigid objects. To achieve real-time performance, the algorithms take advantage of several characteristics of surgical training: (1) visual realism is more important than accurate, patient-specific simulation; (2) most tissue deformations are local; (3) human-body tissues are well damped; and (4) surgical instruments have relatively slow motions. Each key algorithm is described in detail and quantitative performance-evaluation results are given. The specific application considered in this paper is microsurgery, in which the user repairs a virtual severed blood vessel using forceps and a suture (micro-anastomosis). Microsurgery makes it possible to demonstrate several facets of the simulation algorithms, including the deformations of the blood vessel and the suture, and the collisions and interactions between the vessel, the forceps, and the suture. Validation of the overall microsurgery system is based on subjective analysis of the simulation’s visual realism by different users.

Introduction

As computer power and graphics capabilities continue to increase, there is growing interest in surgical simulation as a technique to enhance surgeons’ training. Such training currently requires cadavers or laboratory animals. A computer simulation option could reduce costs and allay ethical concerns, while possibly decreasing training time and providing better feedback to the trainees. However, for surgical simulation to be useful it must be realistic with respect to tissue deformation, tool interactions, visual rendering, and real-time response. This paper describes a microsurgery training system based on novel computer simulation techniques. The system allows a user to interact with models of deformable tissues using real surgical instruments mounted on trackers. It generates a graphic rendering of the tissue deformations in real-time.

Real-time simulation of deformable objects is needed in many areas of graphic animation, for example to generate cloth motions in animated movies or video games, to provide realistic facial animation for digital actors, and to deform soft tissues in surgical simulations. Deformable objects raise a complex combination of issues ranging from estimating mechanical parameters, to solving large systems of differential equations, to detecting collisions, to modeling responses to collisions. See (House and Breen, 2000) for problems and techniques in cloth modeling and (Delingette, 1998) for issues arising in surgical simulation. Many issues still lack adequate solutions, especially when simulation must be real-time.

Here we focus on fast and realistic simulation of tissue and suture, and detecting and processing collisions among rigid and deformable objects. Our main goal is to develop efficient data structures and algorithms that can process large models at a rate compatible with real-time graphic animation (30 Hz). To achieve this goal, we exploit the fact that many deformations are local. By propagating forces in a carefully ordered fashion through an elastic mass–spring mesh, we effectively limit the computations to the portions of objects that undergo significant deformations. To accelerate collision detection, we pre-compute hierarchical representations for all objects in the scene; when objects are being deformed, we only update those parts of the hierarchies that need to be modified.

We have used these algorithms, along with other techniques, to build a system aimed toward microsurgical training. Microsurgery is a well-established surgical field which involves the repair of approximately 1 mm vessels and nerves under an operating microscope. It is a necessity in many reconstructive procedures, including the successful reattachment of severed digits. Using a forceps, the surgeon maneuvers a suture (needle and thread) through the two ends of a severed vessel and ties several knots to stitch the two ends together. The two parts of the vessel undergo deformations caused by their interactions with the suture and the forceps. The surgeon receives only visual feedback, as the vessel is too small to produce any perceptible reaction force. Microsurgeons typically acquire their initial skills through months of practice in an animal lab, at which point they still require months of supervision in the operating room. Without practice, these skills can quickly degrade. The need for such a suturing simulation has been previously addressed in (Delingette, 1998), and a performance study in (O’Toole et al., 1999) discusses the validity of using such a simulator to develop surgical skill, although it does not provide technical details of the actual simulation. Other aspects of suture simulation are discussed in (Kühnapfel et al., 2000).

The main contributions of this paper are the algorithmic tools that we propose for simulating deformable tissue in real-time. We provide extensive quantitative analysis of the performance of these tools. On the other hand, the experimental microsurgery system has only been validated through subjective visual analysis by several users.

Section 2 describes an overview of our simulation system. 3 Computation of object deformations, 4 Simulation of the suture present our simulation and collision-detection algorithms, with specific references to the microsurgery simulator. Section 6 discusses current and future work.

Section snippets

System overview

Our software system includes a deformable object simulator, a tool simulator, and a collision detection module. A graphics display allows multiple objects to be rendered from a 3D virtual world onto the screen at 60 Hz or higher. The user has complete control of the view, and may use stereo glasses for true binocular depth perception (rendered at 30 Hz or higher per eye). The positions of the objects are read from the deformable object and tool simulators before each screen refresh.

Deformable

Relation to previous work

Research on modeling deformable objects has increased dramatically in the past few years. Most proposed 3D models/techniques fall into two broad categories, mass–spring meshes and finite elements.

A mass–spring mesh is a set of point masses connected by elastic links. It represents the tissue geometry and is used to discretize the equations of motion. At each node N an equation defines the force exerted on N by the nodes to which N is connected. Mass–spring models have been used in facial

Simulation of the suture

The suture is deformable but not elastic, so the above deformation techniques based on mass–spring models are not applicable. Instead, the suture behaves as a needle and thread, which can stretch minimally if at all, and has a free-form shape which is affected by gravity and direct contacts. To achieve realistic deformation, we model the suture as an articulated object: 200 short straight links are sequentially connected at nodes which act as spherical joints. The joints allow two degrees of

Collisions and interactions

Interaction with virtual tools is a necessary component of any realistic simulator, and has attracted recent attention (Basdogan, 2001). Almost all object interactions depend at some level on collision detection (Lin and Gottschalk, 1998). Grabbing is achieved by finding nodes colliding with the tip of the grabbing object (e.g. forceps). Piercing the vessel requires finding a collision between the needle edges and a vessel face. Draping the suture around another object also involves edge to

Conclusion and future work

We have designed new fast algorithms for simulating the deformations of soft objects and detecting collisions among deforming and rigid objects. These algorithms take advantage of several characteristics of surgical training: (1) visual realism is more important than accurate, patient-specific simulation; (2) most deformations are local; (3) human-body tissues are well damped; and (4) surgical instruments have relatively slow motions. Our simulator exploits these characteristics to solve

Acknowledgements

This research was conducted in the Stanford–NASA Biocomputation Center. It was supported by grants from the National Aeronautics and Space Administration (NAS-NCC2-1010), the National Science Foundation (IIS-9907060), and the NIH National Libraries of Medicine (NLM-3506). It has also benefited from equipment gifts made by Sun Microsystems and Intel. This paper was improved by discussions with Cynthia Bruyns and Frédéric Mazzella. Fig. 3 was generated by Cynthia Bruyns using the deformable

References (34)

  • J.D Cohen et al.

    I-COLLIDE: An interaction and exact collision detection system for large-scale environments

  • S Cotin et al.

    Real-time elastic deformations of soft tissues for surgery simulation

    IEEE Transactions on Visualization and Computer Graphics

    (1999)
  • S Cotin et al.

    A hybrid elastic model allowing real-time cutting, deformations and force-feedback for surgery training and simulation

    The Visual Computer

    (2000)
  • H Delingette

    Towards realistic soft tissue modeling in medical simulation

  • S Gottschalk et al.

    OBB-tree: A hierarchical structure for rapid interference detection

  • D.H House et al.

    Cloth Modeling and Animation

    (2000)
  • A Joukhadar et al.

    Dynamic simulation: model, basic algorithms, and optimization

  • Cited by (0)

    Electronic Annexes available. See www.elsevier.com/locate/media.

    This paper is based on earlier work appearing in (Brown et al., 2001a, Brown et al., 2001b.

    View full text