This PhD thesis describes a study which investigates the possibility of implementing the Delft Virtual Window System [DVWS, App. I, Smets 1995b, Overbeeke et al. 1988] to achieve 3D Laparoscopy, or to phrase it better, to support the surgeon with spatial information. The DVWS is a technical implementation of viewpoint parallax. It provides spatial information by linking the motions of a camera to the head movements of the observer. The observer controlling the camera obtains spatial information about the object displayed on the monitor. Moreover, he can relate the object’s size to his own movements and thus to himself. Spatial information is obtained by coupling perception and action. You can download the full document here.
The aim this research is the implementation of the DVWS for laparoscopy. Laparoscopy is a minimal invasive surgical technique during which the patient is treated through small ‘key-holes’ (trocars) in the abdomen wall. For obtaining visual information a laparoscope is used, which is basically a tube with a built-in lens system for observation, and with built-in light guides for illumination. On the outer end of the laparoscope a camera is mounted, the image from which is presented on a monitor. Spatial information is limited for a number of reasons. For example, laparoscopes have an integrated light source, which reduces the visibility of shading and therefore flattens the image. Also, the laparoscope is not held or directed by the surgeon, because he needs both hands to operate, but by an assistant. Implementing the DVWS will restore the surgeon’s direct control over the laparoscope and allow him to obtain spatial information through exploration. What kind of information is needed depends on the task to be performed.
Laparoscopy consists of two types of tasks (Ch. 3); observation and manipulation tasks. Both types differ in the spatial information they require. Observation tasks require the perception of the spatial structure, for example, to identify the various tissues. Manipulation tasks, however, also require the perception of how the layout is related to the observer. This is needed, for example, to move from one location to another. The information needed for both tasks (about the spatial layout and about how this layout is related to the surgeon) can be obtained through explorative movements.
For both tasks an implementation of the DVWS seems feasible. However, two major problems arise. First, the feasibility of the DVWS depends on whether or not the surgeon will move. Whereas observation tasks invite movements, manipulation tasks do not (Ch. 4). For example, when manipulation requires a high degree of manual stability, the surgeon will sit as still as possible. However, when little hand stability is required, it is found that the ability to explore can be an advantage (Ch. 6). Moreover, the assistant holding the laparoscope was found to influence the surgeon’s performance of a manipula- tion task. An implementation of the DVWS should also support the laparo- scope.
Secondly, implementing the DVWS (viewpoint parallax) is technically complex, and therefore balances between perceptual criteria (what kind of information has to be provided) and technical criteria (what kind of informa- tion can be provided). A simpler technical implementation is shadow parallax, i.e., the motions of the light source are linked to the movements of the observer (Ch. 5). Shadow parallax provides similar perceptual information (variation in shading and shifting shadows) but has the advantage that a moving light source can be implemented by two stationary light sources of which the inten- sity balance varies. Its implementation thus only requires two separate light guides. Shadow parallax was found to be applicable for detecting differences in heights (industrial application, Ch. 5) but, it was found to provide insuffi- cient information for application for laparoscopy (Ch. 6). For laparoscopy, viewpoint parallax has the advantage.
To arrive at a simple technical implementation of viewpoint parallax, the large viewing angle (fish eye) that laparoscopes commonly have was made use of. This fish-eye makes it possible to direct the laparoscope away from the area of interest, while maintaining the area within vision. Directing the laparoscope will translate the tip of the laparoscope and, as a result, the area of interest is will be observed from a different point of observation
The DVWS was implemented into a working prototype, the Circular Guide (Ch. 7). The Circular Guide links the head movements of the surgeon to the rotation of the laparoscope around the point where it enters the abdomen. The Circular Guide can be used in combination with a common laparoscope. It was designed such that the surgeon can easily control his point of observation either through head movements or manually. Additionally, a mounting device was designed to simplify (re-) installing the Circular Guide.
Both the Circular Guide and the mounting device were tried in a practical setting. They allowed the surgeon to perform an operation without needing assis- tance for the manipulation of the laparoscope. The circular guide was found to work properly, but the mounting device was found to need some adjustments. Overall, the surgeon was enthusiastic about the regained control over his visual information.