SE4D03 Computer User Interfaces

Transcription

1 SE4D03 Computer User Interfaces The Science of Visualization continued Visual Attention Attention comes in two types of considerations: 1. Low-level as considered below, and 2. High-level as considered later, where problem solving is most important Consider the eyeball a searchlight, sweeping the visual world under the guidance of human cognitive centres, where a snapshot is taken upon every fixation of the eye. Each snapshot will take a parse of between 4 and 12 items per fixation and is held in a visual buffer, called the iconic buffer which holds the snapshot for 0.1 or 0.2 seconds as it is read into short term memory. The eye can take on two types of attention spaces: 1. Detailed narrow focus of information, which concerns the image concentration about the fovea where vision detail is most acute, and 2. Generalized pattern sensing where a larger space can be absorbed but with much less detail than in (1). se4d03 slide L30-1

2 The Science of Visualization -- I Visual Attention cont'd Eye Movements moving the eyes causes different parts of the visual environment to be imaged on the high-resolution of the fovea; for instance as you read a line of text in a book, your eye makes between two and five jerky movements which are called saccades, per second. In visual searching, the eye moves from fixation to fixation, taking about ms to extract the information and from 20 to 100 ms for the saccade. Measured maximum saccade rates have been clocked at 900 degrees/second If the visual field contains moving objects, the eye can lock-on and track the motion; this is called smooth pursuit eye movement. If the visual field contains movement towards or away from the eye, the eye will accommodate this by either converging as the motion comes closer or diverge if it moves further away. This is called vergence movement, which can be either smooth or saccadic. This accommodation to refocus can take about 200 ms. Note that the refocus and eye movement are neurologically coupled so that virtual reality displays may be fraught with illusion possibilities. se4d03 slide L30-2

3 The Science of Visualization -- II Low-Level Visual Attention A concept has been defined as the Useful Field of View (UFOV) UFOV defines the size of a region from which humans can extract information that is displayed. For instance, if the material displayed is highly dense than the UFOV may be 1-4 degrees of visual angle; however, if it is sparsely populated then the UFOV may be as large as 15 degrees. The UFOV varies roughly with the target density to maintain a constant number of targets in the region of interest. This concept can also aid in understanding the effect of tunnel vision when the operator is under stress. When this happens the UFOV is narrowed so that only the most important information at the centre of the region is processed. In fact, tunnel vision has a more general meaning that when the cognitive load goes up in an individual, the UFOV narrows. Therefore these concepts are important when an interface must attract the user s attention away from normal operations where an immediate response to an emergency situation is required. se4d03 slide L30-3

4 The Science of Visualization -- III Low-Level Visual Attention cont d Preattentive Processing When viewing visual objects, some objects appear to pop-out from their surroundings when viewed. That is, prior to conscious attention, something of an early visual processing occurs which selects those items for examination; this is called preattentive processing. In interfaces that purport to offer information at-a-glance, then arrangement of items must be such as to identify for preattentive processing those objects that are critically important beforehand. Features of objects that can be preattentively processed include: FORM: line orientation, line length, line width, line collinearity, size, curvature, spatial grouping, blur, added marks (graphical annotations), numerosity,. COLOUR: hue and intensity MOTION: flicker and detection of motion. SPATIAL POSITION: 2D position, stereoscopic depth and convex/concave shape from shading Some examples follow on the next slides. se4d03 slide L30-4

5 The Science of Visualization -- IV Low-Level Visual Attention cont d Preattentive Processing All samples at right are preattentive except Juncture & Parallelism se4d03 slide L30-5

6 The Science of Visualization -- V Low-Level Visual Attention cont d Preattentive Processing Therefore are there item dimensions that emphasize preattentiveness property? Generally no, since this depends on the strength of the particular feature and the context. More contrast is good, but more complexity is not, as shown in the bottom two right in the previous slide. Thus it is better to highlight a word than to underline all but the attentive word. In fact, simple numerosity is preattentive but complex numerosity is not. At the lower left is an example where we need to locate a grey square, from amongst black and grey squares and circles. This is NOT preattentive because we must serially search for both squares and grey colour. This is called a conjunctive search. se4d03 slide L30-6

7 The Science of Visualization -- VI Low-Level Visual Attention cont d Preattentive Processing Conjunctive Searching There is a caveat when conjunctive searching is involved and that is using spatial grouping. Above, the upper right cluster is not preattentive when looking for the grey circle but in the lower right, the grey ellipse is preattentively processed. Highlighting Given the use of computer graphics, there is other methods of highlighting. One example is the blur (and also the grey-over or disable) common in graphical operating systems. The example at the lower left is the blur (a generalization of the fish-eye lens effect). However, here in visualization land, this is termed semantic depth of field variation. se4d03 slide L30-7

8 GUI du Jour -- I Navigation Considerations Selecting change brings up Computer Name Changes which also does workgroup name changes! se4d03 slide L30-8

9 Navigation Considerations cont d Insert Function brings up Paste Function window GUI du Jour -- II NOTE: insert means input, paste implies a previous cut se4d03 slide L30-9

10 GUI du Jour -- III Navigation Considerations cont d Adobe main window (left) allows users to select <New Group>, but the window brought up (right) is titled Edit Group se4d03 slide L30-10

11 GUI du Jour -- IV Navigation Considerations cont d Selecting Mail Folders in upper windows brings up the Manage Folders window se4d03 slide L30-11

12 GUI du Jour -- V Navigation Headers Upper is Subscribers Lower is Newsletters Highlighting se4d03 slide L30-12