TY  - JOUR
T1  - Using ImgSeek to Support Pharmacy Students for Pill Identification
AU - Lertnattee, Verayuth AU - Benjawan, Sirikanlaya 
JO  - The Social Sciences
VL  - 12
IS  - 1
SP  - 54
EP  - 59
PY  - 2017
DA  - 2001/08/19
SN  - 1818-5800
DO  - sscience.2017.54.59
UR  - https://makhillpublications.co/view-article.php?doi=sscience.2017.54.59
KW  - Pill identification
KW  -imgseek
KW  -physical characteristics
KW  -image processing
KW  -confidence
AB  - The number of tablets and capsules has been increasing very fast. It is difficult to identify all of them
correctly. To support pharmacy students recognize more medicines, we need some tools that have ability of
searching by images. The purpose of this study was to investigate the usability of the imgSeek, image
processing open source software, for pills (tablets and capsules) identification. The photographs from both
sides, i.e., front and back of 100 pairs of different pills were used for investigated. Two experiments were done,
i.e., difference on a pair of identical pills and difference on patterns of physical characteristics. For patterns of
physical characteristics, differences of color patterns, imprints and scorings are considered. The percentage
of similarity and rank of the correct images were collected. The mean/median differences of these values were
evaluated by statistical methods with 95% confidence level. The results from both of the percentage of similarity
and rank for difference on a pair of identical pills and difference on patterns of physical characteristics were high
significantly different (p<0.01). From the results, imgSeek could be applied for tablets and capsules
identification. Students should be informed that a long ranked list of pills might be considered especially, a pill
with simple physical characteristics.
ER  -