 |
 |
 |
 |
 |
 |
 |
 |
|
Eye Control: a SciArt 2001 project |
| 1.
Background |
|
Can painters who draw from life assist radiologists searching for signs of lung disease in X-ray or CT scans? Can research into the movements of a painter's eyes be applied to detect fatigue in surgeons performing lengthy operations? How legibly can one sign one's name by eye, and is this a good test for eye control? These are some of the questions that Dr John Tchalenko, head of Drawing and Cognition at Camberwell College of Arts, and Dr Guang-Zhong Yang, head of Visual Information Processing at Imperial College will be examining in their recently awarded SciArt project - EYE CONTROL. The first paper to arise from this project, Quantitative Analysis of Eye Control in Surgical Skill Assessment, presented at the 11th European Conference on Eye Movements in August 2001 in Turku, Finland, may be see here. |
|
 |
| 2.
Description |
|
NB - The figures included in this text are low resolution JPEG and GIF files. You may download high resolution Tiff files by contacting either G Z Yang or J Tchalenko at the e-mail addresses given below.
|
|
Two years ago, John co-ordinated the Painter's Eye exhibition at the National Portrait Gallery, showing the results of a detailed investigation on how a painter draws from life. Since then, he started using an unusual instrument, the Eyemouse, to find out how much control artists and student artists have over their voluntary eye movements. With this instrument the computer cursor is controlled with the eyes, allowing the operator to follow a curve, draw a picture or sign a name on the monitor with the eyes alone (FIGURE 1). |
 Fig 1. Drawing with the Eyemouse. The background image of the operator's eye is optional. |
|
Most of us are not very good at such tasks. In normal life, our eyes move in rapid jerks (or "saccades") from one point to another about 150 times per minute. We can decide to change our glance from one part of a scene to another but the actual path our eyes take is something we are not aware of and have very little control over. This is what all the experts in Visual Perception had predicted, but John found that there were some remarkable exceptions, namely amongst painters who drew regularly from life. They seemed to have strong control over their eyes and could move them very slowly and smoothly to write their names or draw a simple picture on a blank screen (FIGURE 2). |
 Fig 2. Writing with the eye. The painter Humphrey Ocean has much better eye control than John Tchalenko. |
|
The Visual Information Processing unit at Imperial College uses state-of-the-art eyetrackers to provide the precision required for a complete analysis of eye movements. When John first visited Guang-Zhong's team, he was struck by the similarity of some of the questions being investigated there with his own work on painters. A viewer in an art gallery seeing a painting for the first time has a complex eye scan path in search of answers to questions raised by the picture (FIGURE 3). |
 Fig 3. Seeing a painting. Scan path of subject B.R.'s eye as he is seeing Double Portrait by Humphrey Ocean for the first time. B.R. later informed us that he could not understand how the camera operator in the picture was looking through the viewfinder. |
|
Similarly, looking at a CT image of a lung, the novice's eye follows a complex and disorderly path in its search for the darker textured patches indicating diseased areas (FIGURE 4a). |
 Fig 4a. Novice reading a CT image of a diseased lung. The eye has trouble in distinguishing the relevant darker patches. |
|
But Guang-Zhong also observed that the consultant radiologist had a much clearer and economical search pattern than the novice (FIGURE 4b). |
 Fig4b. Specialist consultant radiologist seeing the same image as in Fig 4a. The search pattern is much clearer and more precise than for the novice. |
|
John had also observed that when drawing portraits from life, the professional painter's eye movements between model and drawing were simpler and more precise than those of novices drawing for the first time. The indications were therefore that both in drawing and medicine, training and practice had resulted in controlled patterns of eye movements optimized for the task in hand. Could this eye control factor be assessed objectively and used for training and monitoring purposes? Like in portrait drawing, some medical tasks, such as surgery, require fine co-ordinated eye-hand control that can be affected by fatigue and stress. Could a quick and easy eye control test be devised to monitor such effects during lengthy surgical operations? The sciart project grew out of these shared questions: it was decided that both artists and medics would be tested on the eyemouse and on the eyetracker to explore the potential applications of this approach.
|
|
Meanwhile, back in Camberwell, having gone through his test routine, the Spanish painter Manuel Franquelo is "doodling" with the eyemouse. He is exploring his idea of using the natural jerkiness of eye movements to produce images by an overabundance of lines (FIGURE 5). |
 Fig 5. The painter Manuel Franquelo trying out his eye at "Free-eye drawing". Scientists only partly understand why the eye is in constant motion. |
|
Actually, this is also of interest to us: despite over 50 years and hundreds of studies, scientists have as yet to find a use for the smallest of these sharp movements they call "microsaccades". They seem to have no role whatsoever in our ability to see. Manuel may be on to something. Painters, after all, are the real experts in Visual Perception! |
|
|
| 3.
Contacts |
| Dr
Guang-Zhong Yang: gzy@doc.ic.ac.uk Dr John Tchalenko: j-t@dircon.co.uk |
| 4.
Links |
|
|
|
|
Drawing & Cognition Project leader: Dr
John Tchalenko
E-mail: j-t@dircon.co.uk Phone: (+44) 020 7514 6364 See also: The Eye Mouse Project 
Webmaster: i.warren@linst.ac.uk Last update: |
 |
|
 |
 |
 |