Comments:
Summary:
This paper presents the need to manipulate objects presented on large high resolution displays from both close up and a far. The pointing device must be able to transition smoothly through this range to be effective. Under these requirements traditional devices are not adequate. A more naturalistic approach of using hand gestures is presented. The vocabulary of hand gesture is described along with some of the challenges in their implementation. For example the calibration of the air tap posed some difficulties as there is no lower limit to the movement. In contrast the thumb trigger although providing tactile feedback proved uncomfortable and tiring. Both visual and auditory feedback signals were presented to aid the user. Additionally a graded ambiguous posture visualization was created which greatly facilitated learning. Three types of pointing were explored. Ray casting using the index finger and a dynamic recursive lowpass filter. Relative pointing and clutching using a dynamic lowpass filter to eliminate jitter in combination with relative measurements taken from a starting position. A simple and effective use of cursor orientation was used to provide feedback when clutching. Finally RayToRelative pointing with simultaneous recalibration.
The goals of the study were to compare the task completion time, error rate, and recalibration characteristics. The user was required to hold their hand in a predetermined position for two seconds before the cursor and first target would appear. The user was required to select it before the next trial. RayCasting was found to be faster where clutching would have been required or when selecting large targets but its high rate of error prevented it from being practical. No significant difference was found between Relative and RayToRelative techniques.
Discussion:
A very interesting paper presenting a minimalistic solution to maximize the potential of very large high resolution display usage.
The work is clearly presented and thoroughly executed and tested.
A point which is covered in the article itself is the elimination of hand markers, to make the execution solely in the hand of the user.
I really like their solution, as I find it very elegant and natural, since we use similar hand motions already in many environments, especially pointing at big displays.
ReplyDeleteElimination of the hand markers would make this much better, especially since the markers looked rather large. I wonder if the markers have any effect on the interaction due to their size.
I didn't mind too much about the usage of hand markers, since I felt the paper was more on establishing the correct techniques. Technology hasn't matured enough yet to have accurate non-invasive hand gesture recognition yet, and so hand markers was a necessity in order to perform the studies in this paper. The alternative would have been to simply not have the study until the technology is ready, or spend their time building the technology before they can run their study, but that would just delay the study that is independent of the technology.
ReplyDelete