Using Kinect for 2D and 3D pointing tasks: performance evaluation

TY - GEN

T1 - Using Kinect for 2D and 3D pointing tasks

T2 - 15th International Conference on Human-Computer Interaction

AU - Pino, Alexandros

AU - Tzemis, Evangelos

AU - Ioannou, Nikolaos

AU - Kouroupetroglou, Georgios

N1 - Conference code: 15

PY - 2013

Y1 - 2013

N2 - We present a study to comparatively evaluate the performance of computer-based 2D and 3D pointing tasks. In our experiments, based on the ISO 9241-9 standard methodology, a Microsoft Kinect device and a mouse were used by seven participants. For the 3D experiments we introduced a novel experiment layout, supplementing the ISO. We examine the pointing devices' conformance to Fitts' law and we measure a number of extra parameters that describe more accurately the cursor movement trajectories. Throughput, measured in bits per second is the most important performance measure. For the 2D tasks using Microsoft Kinect, Throughput is almost 39% lower than using the mouse, Target Re-Entry is 10 times up and Missed Clicks count is almost 50% higher. However, for the 3D tasks the mouse has a 9% lower Throughput than the Kinect, while Target Re-Entry and Missed Clicks are almost identical. Our results are also compared to older studies, and we finally show that the Kinect, operated by the user's hand and voice, is a suitable and effective input method for pointing and clicking, especially in 3D tasks.

AB - We present a study to comparatively evaluate the performance of computer-based 2D and 3D pointing tasks. In our experiments, based on the ISO 9241-9 standard methodology, a Microsoft Kinect device and a mouse were used by seven participants. For the 3D experiments we introduced a novel experiment layout, supplementing the ISO. We examine the pointing devices' conformance to Fitts' law and we measure a number of extra parameters that describe more accurately the cursor movement trajectories. Throughput, measured in bits per second is the most important performance measure. For the 2D tasks using Microsoft Kinect, Throughput is almost 39% lower than using the mouse, Target Re-Entry is 10 times up and Missed Clicks count is almost 50% higher. However, for the 3D tasks the mouse has a 9% lower Throughput than the Kinect, while Target Re-Entry and Missed Clicks are almost identical. Our results are also compared to older studies, and we finally show that the Kinect, operated by the user's hand and voice, is a suitable and effective input method for pointing and clicking, especially in 3D tasks.

U2 - 10.1007/978-3-642-39330-3_38

DO - 10.1007/978-3-642-39330-3_38

M3 - Article in proceedings

SN - 978-3-642-39329-7

T3 - Lecture notes in computer science

SP - 358

EP - 367

BT - Human-Computer Interaction. Interaction Modalities and Techniques

A2 - Kurosu, Masaaki

PB - Springer

Y2 - 21 July 2013 through 26 July 2013

ER -