3D vision

作者群

摘要

The description of 3D vision covers all aspects of 3D vision relating to the operating room.
The technical key steps of the chapter are presented in a step by step way: historical overview, physiology of 3D vision, visual conditions, available material, advantages/disadvantages.

媒體類型

刊物

2005-09

普通的

最愛

音訊

數位出版

WeBSurg.com, Sept 2005;5(09).
URL: http://www.websurg.com/doi-ot02en309a.htm

3D vision

1. Introduction

Three-dimensional (3D) vision plays an important role in the performance of precise surgical manipulation. Whereas surgeons have complete 3D vision in open surgery, when performing laparoscopic procedures they are denied the stereoscopic component of 3D vision and have to learn to work with a two-dimensional (2D) view of a surgical field. For the past few years, manufacturers have been developing different solutions to reproduce depth perception by generating stereoscopic cues. Although it has been clearly demonstrated that stereoscopic vision can improve surgical performance, current technology is limited by the projection systems used to present the image of the surgical field to the surgeon.

2. Historical overview

In 1904, L. and H. Loewenstein were granted a German patent for their stereocystoscope.
Unfortunately, optical technology of the time did not allow the full realization of the potential of this invention. It was only in the 1970s that R. Wolf invented a good quality stereo-endoscope with new lenses that offered much better optical properties. Despite the interest shown in this new equipment, it was not used as widely as expected, and urologists continued to work with monoscopic endoscopes. With the extension of endoscopic surgery to other specialties requiring more complex manoeuvres, an increasing interest for 3D vision has developed. At the present time, stereoscopic systems are routinely used in ophthalmic and transanal surgery. However, in general laparoscopic surgery, the technology is not sufficiently advanced to provide systems that can compete with the visual quality of good monoscopic endoscopes.

3. Physiology of 3D vision

• Binocular vision

• Vision

Humans, most primates, birds of prey and cats have binocular stereoscopic vision. Because both eyes are placed at the front of the skull and look in the same direction, their fields of view overlap.
In order not to have a double image, the brain controls both eyes in such way that they focus exactly on the same point (point of fixation) in a coordinated manner (convergence and accommodation).

• Stereoscopic vision

If the object is placed closer than 9 meters, then the retinal projection of the image is different for each eye. This deviation is called lateral disparity and is due to the fact that each eye perceives images at a slightly different angle. The brain interprets this disparity as a sense of depth of the observed object producing stereoscopic vision. At greater distances the sense of stereoscopic vision is lost because the eyes’ visual axes are essentially parallel and the brain perceives no difference between the images in each eye. For distant objects the brain uses other available visual cues to generate a sense of depth. We may note that many other animals, such as pigeons, rabbits and most fish, have panoramic vision. Their eyes are placed on the sides of their head, allowing for a very wide field of vision with little overlap between the two eyes. This prevents them from having stereoscopic vision.

• Image interpretation

The brain utilizes a knowledge of how our world is constructed to deduce the relative position of objects in the visual field. Examples:
- An object covering another object is perceived as closer to the observer.
- Shadows give an indication of the relationship between an object and the structure upon which its shadow falls. For example an object appears in contact with a surface when it touches its shadow on that surface.
- The relative size of an object is used to estimate its distance if its real size is known.

• Parallax movement

As objects in a visual field move relative to the observer, the brain is able to determine relative distances as a function of how different objects in the field move relative to each other. For example when looking out of a window on a train objects close to the observer seem to move backwards relative to the horizon much more quickly than objects, which are further away.

• Visuomotor signals

• Accommodation

The brain uses information on how the eyeballs and ocular lenses are being adjusted to give clues about the distance of an object.
The ocular lens contracts to obtain a clear retinal image of an object close to the observer.

• Convergence

To focus on an object close to the observer the eyeballs turn inwards (convergence). The brain interprets this convergence as an indication of how close the object is to the observer.
Under normal conditions, human beings are capable of using all these depth cues to achieve accurate hand-eye coordination. It is important to note that stereovision is only one component of 3D vision and that even in its absence there is still a sense of depth perception. Indeed, persons suffering from amblyopia will use the other sources of depth information described above to create a mental 3D reconstruction of their environment allowing them to achieve remarkably precise depth perception.

Unfortunately, in standard laparoscopic surgery, most of this information is lacking, requiring the surgeon to learn how to manoeuvre with 2D vision. This handicap makes it necessary for the surgeon to utilize mostly haptic information to generate a sense of spatial awareness making manipulation slower, less precise and less safe.

• Conventional laparoscopy

• Drawbacks

Conventional laparoscopy has the following drawbacks:
- Image blurred by fogging or dirt:
Unlike the human eye that is regularly washed by blinking and tears, the presence of a residue on the tip of the endoscope will result in blurry vision and a decrease in the contrast and sharpness.
- Lack of shadow on the endoscopic image:
The light source is projected from the tip of the endoscope by a ring of light around the lens. The resulting image is bright, but lacks shadows. The various anatomical structures appear as if they were flattened under this uniform light.
- Lack of stereovision:
Traditional endoscopes and video cameras are monocular, making stereovision impossible.
- Inappropriate convergence and accommodation:
When a surgeon watches the video monitor, his or her ocular lenses and eyeballs are adapted to the distance of the screen. Since the distance of the screen image bears no relation to the real distance of objects in the laparoscopic field, the brain cannot use information about the degree of lens contraction or accommodation to assess the depth of an object in the operative field of view.

• Lack of parallax movement

When the endoscope is moved, parallax movement is negligible. This is partially due to the limited distance that the endoscope can be moved, but is also related to the limitations on how the laparoscope can be moved (only 4 degrees of freedom rather than 6). In fact, because the laparoscope pivots in the anterior abdominal wall, the surgeon cannot remain focused on an anatomical structure and obtains a different angle of view.
These drawbacks make depth perception and precision movements more complicated for laparoscopic surgeons who must proceed by trial and error with the tips of their instruments to reach their targets. This results in less smooth and less accurate manipulation.

4. Visual conditions

To compensate for the lack of information needed for depth perception, several solutions have been developed that can improve the 3D effect.

Creation of shadows:
Two technical solutions exist:
- a second light source integrated into a trocar, providing additional indirect lighting;
- an endoscope equipped with a second light source directed differently to the axial light source at the endoscope’s tip.
By adding an indirect light source, shadows are projected on the structures being viewed, creating a new sense of depth.

Stereovision:
Two distinct images are sent separately to each eye to create binocular vision.
A stereo-endoscope is composed of two 5 mm optical systems assembled in a single endoscope equipped with two video cameras. Each camera receives the image from one of the imaging channels. The system mimics the human eyes on a smaller scale. The images received by the two cameras have a slightly different point of view. Because the two lenses are separated by only 6 mm on the average, i.e. 1/10 of the distance between human eyes effective stereoscopic vision is produced over a distance of only 10 cm. Another drawback is related to the fact that both optical systems must be contained in a normal sized endoscope (10-12 mm), therefore decreasing the diameter of each channel with a corresponding decrease in the quantity of light transmitted. Inevitably these systems are more costly.

To avoid the drawbacks and high cost of stereo-endoscopes, certain systems use a mono-endoscope equipped with a prism that separates the image into two identical images, which are then sent to two video cameras. The rest of the image processing remains identical to the stereo-endoscopic system. While this system can obtain superior image resolution with higher brightness, the 3D effect is apparent rather than real since the original images were identical.

Digital image processing:
The human eye perceives contrasts better than progressive variations in brightness. An increase in contrast and digital enhancement of outlines by digital image processing helps to clarify the tissue edges, thereby artificially creating a 3D effect.

On older video cameras, this effect was achieved at the expense of the resolution and color rendition. New cameras no longer have this drawback, and the latest generation of digital cameras are expected to offer advanced, high performance video processing, capable of artificially creating 3-D effect without loss of quality.

Improvement in the resolution of the camera-monitor system:
The better the resolution of an image is, the more visible details there are. This can help to increase 3D perception. Current video systems have a resolution of about 600 horizontal lines for the PAL standard (about 500 lines for NTSC), which is considerably less than the resolution perceived by the human eye. A high definition system, such as HDTV (High Definition TeleVision) with 1250 horizontal lines per image for PAL (1050 for NTSC) can improve 3D perception. At the present time, these types of systems are expensive and are rarely used in digestive surgery.

5. Available material

• Creation of shadows

• Option 1

Several technologies have been developed to reproduce a 3D image. The following list is not exhaustive.
Creation of shadows
- second light source integrated into the trocar;

• Option 2

- endoscope equipped with a second light source: ‘’Shadow Telescope’’ of MGB.
This is an economical way to create a sense of depth, but it is rarely used.

• Stereovision: image creation

Various systems have been developed to create 3D images based on the principle of stereovision.
- True stereo-endoscopes with two optical systems (i.e. two imaging channels) and two video cameras. Each camera receives the image from one imaging channel.
- Mono-endoscopes equipped with a prism that separates the image into two images that are sent to two video cameras.
In order for the brain to create a 3D image, a separate image must be projected on to each eye. Three projection systems exist.

• Stereovision: image display

• Head mount

Head mount with separate integrated screens: each image is displayed on a small, color LCD screen placed just in front of each eye. The images are small and the resolution of LCD screens is limited, reducing the quality of the image. Other drawbacks include the fact that the surgeon is visually cut off from the rest of the operating room, and that a head mount must be worn.

• Active glasses

Active shuttered glasses: each image (right and left) is alternately projected on a monitor with a refresh rate of 100 Hz for a PAL system (120 Hz for NTSC). An infrared transmitter is used to beam a signal to synchronize the refresh rate of the monitor with the crystal liquid glasses worn by the surgeon, which alternate from a transparent to an opaque state in such way that each image is visible only to the appropriate eye. Each eye receives an effective image refresh rate of 50 Hz for PAL (60Hz for NTSC). Flickering is not detectable and eye-strain is avoided. The brain then reconstructs a 3D image from the two different images. Liquid crystal glasses have the disadvantage of causing visual fatigue.

• Passive glasses

Passive polarizing glasses: again the two images are displayed on a single monitor at an alternating rate of 100 Hz. An active polarizing filter in front of the monitor polarizes the images first vertically and then horizontally. The surgeon wears passive polarizing glasses with one lens polarizing vertically the other horizontally so that each eye receives only one of the two images. A decrease in image brightness of about 50% results in a dull image that is lacking in contrast. The system also causes visual fatigue.
All of these systems provide 3D effects, more or less successfully. Unfortunately, the quality of the image reproduction remains inferior to that observed with 2D systems.

• Digital image processing

All of today’s video cameras can achieve an increase in the contrast and a digital enhancement of the outlines of the image, in order to distinguish the planes better and artificially create a sense of depth.
The newest digital cameras offer multiple image processing features that improve the quality of the color rendition and the 3D effect. The main advantage of digital technology is its capacity to store information without any deterioration.

Thanks to digital technology, video cameras can now produce high quality images, closer to reality and can be adapted to different situations encountered during surgery (preset parameters that can be selected according to the type of surgery and endoscope). It is possible to process the image to obtain accurate colors, similar to the colors that would be perceived during open surgery and to limit image dazzling caused by diffraction from structures close to the camera, while maintaining good, overall brightness and increasing the contrast in shadowed areas.

6. Advantages/disadvantages

The technologies that make 3D imaging possible are continually improving. The advantage of these systems has been demonstrated in experimental conditions, but has yet to be proven in the clinical setting. The quality of currently available 3D projection systems remains inferior to the latest generation of 2D video cameras. The present 3D systems lead to visual fatigue and a considerable, additional cost compared to standard equipment.

Until studies have been carried out that show a superiority of these systems in the clinical setting, they will remain limited to a few specific applications (e.g. coronary and tubal microsutures).

7. Conclusions

While the loss of 3D vision in laparoscopy is experienced as an additional difficulty for surgeons at the beginning of their practice, as their expertise increases, they are able to find new spatial landmarks. Moreover, continual progress in digital video systems has improved depth perception by digital image processing. In comparison, current stereoscopic systems still cause significant visual fatigue and their complexity and added cost are not currently justified by their proven advantages. Technological advances resulting in well tolerated, high image quality stereoscopic systems could result in their routine use in the future.

使用者名稱
密碼
	記得我 (必須啟動Cookies設定)