C. Stephanidis (Ed.): Posters, Part I, HCII 2011, CCIS 173, pp. 383–388, 2011.
© Springer-Verlag Berlin Heidelberg 2011
Helping Hands versus ERSP Vision:
Comparing Object Recognition Technologies for the
Visually Impaired
Marc A. Lawson
1
, Ellen Yi-Luen Do
1
, James R. Marston
2
, and David A. Ross
2
1
College of Computing, Georgia Institute of Technology, Atlanta, GA 30332 USA
2
Atlanta Vision Loss Center, 1670 Clairmont Road, Decatur, GA 30033 USA
{marc.lawson,ellendo}@gatech.edu
jim.marston@gmail.com, ross0128@bellsouth.net
Abstract. A major challenge for people with vision impairments ranging from
severely low visual acuity to no light perception (NLP) is identifying or
distinguishing the difference between objects of similar size and shape. For
many of these individuals, locating and identifying specific objects can be an
arduous task. This paper explores the design and evaluation of the “Helping
Hand”: A radio frequency identification (RFID) glove that audibly identifies
tagged objects. In this paper we describe the design of a wearable RFID
apparatus used for object identification. We evaluated the effectiveness of the
glove by conducting a three-arm randomized controlled study. In our
experiment, we compare a baseline (no assistive device), RFID (Helping Hand)
and computer vision (ERSP Vision Software) in identifying common household
objects. We also administered a questionnaire to obtain subjective data about
the usability experience of the participants. Our experimental results show a
reduction in the amount of time required to identify objects when using the
Helping Hand glove versus the other two methods.
Keywords: object recognition, RFID, visual impairment, assistive technology,
and computer vision.
1 Introduction
Over 314 million people worldwide are visually impaired, of whom 45 million are
totally blind [1]. Presently, in the United States there are approximately 20 million
people with visual impairments [2]. There are several different gradients of visual
acuity – ranging from severely low visual acuity to no light perception (NLP). Many
tests are developed to measure a person’s level of vision. For example, Tumbling eye
charts and the Snellen test measure an individual’s visual acuity against the normal
population curve. Performance scores are ranging from 20/10 to 20/200, where a
score of 20/20 represents “normal vision” or what the average person can see at a
distance of 20 feet. People with extremely low visual acuity have what is called a low
vision evaluation. These evaluations are more comprehensive and measure a patient’s
ability to count fingers, observe motion or perceive light from varied distances [3].