Case Studies - Measuring the User Experience: Collecting, Analyzing, and Presenting Usability Metrics - page 245

operator-assisted process. US-VISIT is considering unassisted biometric cap-
ture for specific applications. But ensuring acceptable quality images requires
that users receive real-time feedback on performance. Many factors influence
image quality, including fingerprint positioning, alignment, and pressure. What
form this informative feedback should take is the challenge for an international
audience.

To address this need, Guan and colleagues (2011) designed an innovative,
cost-efficient, real-time algorithm for fingertip detection, slap/thumb rotation
detection, and finger region intensity estimation that feeds rich information
back to the user instantaneously during the acquisition process by measuring
objective parameters of the image. This study investigates whether such rich,
real-time feedback is enough to enable people to capture their own fingerprints
without the assistance of an operator. A second objective is to investigate if pro-
viding an overlay guide will help people in better positioning their hands for
fingerprint self-capture.

10.2.1 Methodology
EXPERIMENTAL DESIGN
We used a within-subject, single factor design with 80 participants who per-
formed two fingerprint self-capture tasks: one with a fingerprint overlay dis-
played on a monitor to guide them on positioning their fingers during the
capture process and the other task without the overlay. Order of receiving condi-
tions was reversed for half the participants. The dependent variables are:

• Task completion rate —ratio of the number of participants who completed

the self-capture task versus the number who did not complete or com-
pleted the task with assistance

• Errors —number of hand-positioning corrections until an acceptable fin-

gerprint image is recorded

• Quality of the fingerprint image —the NIST fingerprint imaging quality

(NFIQ) scores of the fingerprint images

• Attempt time —time from the moment a participant presents her hand

until the end of the capture

• Task completion time —total time it takes to complete a capture task

• User satisfaction —user's ratings from the post-task questionnaire

PARTICIPANTS
Eighty adults [36 females and 44 males; ages ranging from 22 to 77(mean =
46.5)] were recruited from the general population (Washington, DC, area).
Participants were distributed diversely across education, occupation, and eth-
nicity. Fifty-four participants indicated that they had been fingerprinted before:
18 had prior experience with inked and rolled fingerprinting and 36 did not
indicate the type of their fingerprint experience. All fingerprint experiences were
assisted.