http://dl.acm.org/citation.cfm?id=2207733&CFID=108704629&CFTOKEN=72620817
Richard T. Guy & Khai N. Truong
Richard T. Guy is a PhD student in Computer Science at the University of Toronto. He explores the possibilities of computational imaging for assistive and collaborative interaction.
http://www.cs.toronto.edu/~guy/
Khai N. Truong is an Assistant Professor in the Department of Computer Science at the University of Toronto. His research interests include human-computer interaction and ubiquitous computing.
http://web.cs.toronto.edu/people/profiles/faculty/truong.htm
Summary
Richard T. Guy and Khai N. Truong were doing some work on CrossingGuard which aids visually impaired individuals by giving them greater information of their surroundings at unfamiliar intersections. CrossingGuard takes 9 different categories of information and has 85.5% accuracy derived from them. Many standard GPS devices display only the fastest way for cars regardless of if there are good sidewalks to walk on which would be troublesome for a visually impaired person. Street signs are also a hindrance since they can be hard to read for the individual.
The traditional GPS systems that are marketed toward these impaired individuals distinguish objects such as intersections or sidewalks with sound, wind, or light intensity. These systems were not reliable...most notably the sound category because of rainy days. Water causes stronger echoes which could confuse an individual easier.
CrossingGuard was designed to be sidewalk-to-sidewalk so the person would not get lost as easily. CrossingGuard also has built in touch screen support which different types of taps setting the location or even giving information about the immediate intersection. These functions can more readily assist the individual and reduce downtime. On a side note, this technology works by connecting a series of nodes together which store various bits of information such as street name.
Guy and Truong then grabbed 10 people and tested their CrossingGuard application on them. These participants let the authors know which of the audio indicators were most helpful and which ones were not. Many participants said that this app was most useful when travelling to new areas. It is a huge benefit to use CrossingGuard in new areas because of the detailed information supplied such as: intersection shape and road width. This app lets visually impaired people travel independently and fearlessly.
The work Guy and Truong are performing is novel in the way that they are achieving the end result. They might have greater accuracy in helping the visually impaired navigate through intersections than other products would. Having said that, there are many other products currently being researched and even on the market that aim to do the same thing (GPS for the visually impaired) and have been for a while that CrossingGuard does.
Other papers with similar topics:
http://dl.acm.org/citation.cfm?id=1427016.1427201&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=1378063.1378098&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=1078023.1078188&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=2015035.2015334&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=968876.968932&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=2310658.2310895&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=1639642.1639662&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=2042283.2042392&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
http://dl.acm.org/citation.cfm?id=1852204.1852209&coll=DL&dl=GUIDE&CFID=109434563&CFTOKEN=89409536
This paper compares to the first paper, Crosswatch, a camera phone system, because they both index sidewalks for the physically impaired to aid them in crossing intersections and streets. A few of the papers snap video of the surroundings to guide the pedestrian which is quite similar. They might vary in how they gather the information, but they all have the same idea of guiding the physically impaired pedestrian to their destination.
Evaluation
Guy and Truong evaluated the work nicely. They used mainly quantitative measures because of how many statistics they gave of their patients using the app. They also gave some good qualitative measures on their device through describing it in words. So the participants gave constructive feedback-- this is how they got their qualitative data. I would say they used objective measures instead of subjective because of how much time they put into researching other GPS systems marketed towards these individuals on the market. They also put a lot of time into talking with the patients and obtaining constructive feedback from what they wanted in the app.
Discussion
I thought the work and contribution was novel and necessary. There were other products out there on the market that were trying to do the same thing CrossingGuard does but through different means. I do still think this app was novel in the way they designed the algorithm and might even give them greater accuracy. The evaluation was definitely appropriate since they gave other test cases of other GPS versus their own. They also performed experiments and worked out the flaws in their own design through customer feedback. This experiment was necessary to further facilitate an easier and safer commute for these certain individuals.
