CIS-447-101

Prof. Murray Turoff

6 December 1999

1. INTRODUCTION

A large population of the world is disabled and unable to use the current technologies provided to interface with a computer. "Out of every one hundred persons, ten are disabled. In the United States more than twenty million people have a disability that affects their daily lives." (Murphy) Many people with disabilities have a fully functioning brain and have the mental capacity to be successful in this world, but often their physical disability has hindered their success. In a society that physical strength no longer determines your place in society, the disabled are faced with yet another obstacle to overcome. Computers assist the disabled in succeeding, but currently the input devices and interfaces are not designed to assist the physically challenged. Computer input devices and user interfaces are very well designed to assist a user in producing fast output and customization, but this is only for a normal user and can in fact hinder a disabled user.

Computers have the ability to provide a relief to the disabled and their inability to succeed in society. Not only can computers help a disabled user succeed in the business world, but they also bring a new world of opportunities to socially thrive in a virtual society. Many of us take for granted chat rooms, email, and other communication packages but these are excellent tools for one that cannot speak (Murphy). Likewise, the World Wide Web also brings the opportunity for disabled computer users to learn, shop, and travel without the hassle caused by their disability. The virtual world created by the recent boom of computers and the Internet has given those with disabilities the opportunity to live in a world where they are treated as an equal. "For the first time in their lives, many disabled people find themselves able to belong somewhere, a virtual community where they can be swept along by daily events, as in any other community." (Sussman, 85)

There are many products on the market and many more in the research phase. One cheap alternative to using a mouse is the trackball. A trackball uses less arm motion and is an affordable input device for one with a disability, granted the disability still requires some arm movement. One with limited typing abilities can use an enlarged keyboard or a keyguard. A keyguard is a cover for a standard keyboard that makes it near impossible to input the wrong key. (Port 2000 Newsletter) For those who are blind, there are certain keyboards with Braille on the keys and printers that print in Braille (Murphy).

If a user has control of a part of their body, for instance a finger, they can customize switches to interface into a computer using Morse code. Devices are also able to track eyeball motion to move a mouse cursor on screen. These switches and tracking devices are currently used for disabled people who have few other alternatives. Another method of inputting into a computer is called the "sip and puff" technique. This technology allows the user to input information into a computer with air pressure created by the mouth through a tube. (Port 2000 Newsletter)

There are many forms of input devices that have been developed to specially accommodate the disabled in their struggle with life, but this paper will focus on just a few. The three technologies I wish to focus on are screen readers, speech and gesture recognition, and the use of brain signals. Although some of these have been fairly well developed, none have been perfected and all are still in a research phase.

 

2. SCREEN READERS

The blind are perhaps one of the disabled that get cheated the most. Their disability still allows them to fully function as a human being, but with computers they are unable to see the output provided. As I have said before, there are Braille keyboards and Braille printers that allow the user to print in either Braille or regular text. Although these are helpful, the still do not make it possible for a blind person to interface into a computer since they do not know what is on the screen. It has been many years since the first screen reader was developed and they worked quite well, but this was in the days of the DOS text based systems.

The purpose of a screen reader is to translate what is on the screen into spoken language so that the blind can hear what is being displayed. This works especially well with e-mail, chat rooms, and word processors. Problems with screen readers are with graphical interfaces that may not have text that is easily readable to a screen reader. Another problem is that a screen reader tells what text is being displayed, but it is impossible for a blind person to use the system unless they have been properly trained to use the keyboard commands. The current interfaces are designed to make it easy for a sighted person to find a command in a menu or toolbar, but this is difficult without a screen reader that reads menus and tool tips. This means that either the screen reader has to inform the user when the mouse is over a menu or button, or tab keys have to be provided to allow the user to search through all the possible commands with a verbal reply from the screen reader telling the user what command they are currently on.

Screen readers have trouble with images and framed screens on the Internet. One fix to this is Internet Explorer and Netscape’s option to turn off the images while surfing the Internet, which allows for the screen readers to work properly. Problems are still existent when a webpage uses image maps, since there is no textual representation for this. (Murphy) To get around these problems software has been developed that translates and reads the HTML behind the webpage to assist a blind user, but this is growing more and more difficult as website become more involved with Java Applets, Java Script, XML and other emerging modular technologies. One software product that reads HTML into speech is called "PWWebspeak 1.2" and has been developed by a company called Productivity Works, Inc. (Murphy) Currently the best solution is to consider the blind when creating a website and include a text only and no frames option.

Text readers are a great assistance to the blind; an example of this would be Norman Coombs. Norman has been using a speech reader to read the daily news off of the Internet; in the past he would have to wait a week before the papers were in Braille. He also uses his computer to teach a class via the modem. (Sussman, 85)

3. SPEECH RECOGNITION

Another technology that is being developed quite successfully is speech recognition software. Speech recognition allows users to interact with a computer system with their voice. Speech recognition does not mean that the user needs to be able to speak properly, but only consistently.

Speech recognition "allows the user to simply speak commands into a microphone connected to the computer. For best results the system must be trained to recognize the user’s voice. This means the user does not have to speak clearly or accurately by society’s standard, just consistent with the trained pronunciation." (Port 2000 Newsletter)

For several years there has been attempts to create successful speech recognition to be added to existing systems, but these older trails were error prone. Recently, however, these recognition systems have vastly improved to a point where they will be able to be used for the disabled and possibly for everyone. Speech recognition interfaces make the system more intuitive for a non-technical person to use (Grasso, 304).

Most commonly, it is thought that speech recognition is used for computer interaction, but there is another use to this software. In his article Murphy described a use of computer speech recognition for the hearing impaired. With a microphone, a computer, and the correct software the user can train them self to speak properly. In this described software a blind person can speak a selected word into the microphone and receive output on how well they spoke the word. With this type a system, the user can analyze and improve their speaking abilities (Murphy).

Although for some, speech recognition may be the only affordable solution for interfacing with a computer, it has been found to be too cumbersome for the average user. Grasso, Ebert, and Finin show us how a multi-modal interface is far more productive to a normal user. Using a multimodal interface, including regular input devices, handwriting, and speech recognition, it was shown that users enjoyed the system and were more successful in all areas of the system. This includes faster response time, and fewer errors (Grasso, 319-320).

In order for speech recognition to become successful in the common market it is also necessary for the speech to be natural to the user. The system cannot force the user to speak a certain way, or even use specific sentence structures but instead must be intuitive to the user (Hansen).

 

4. GESTURE RECOGNITION

A fairly new field of research for inputting information into a computer is the use of gesture recognition. With this new technology it is possible to read sign language into the computer. Another use of this could be for an alternative to a mouse, or for use in three-dimensional applications.

"Sign language is the usual method of communication for the hearing-impaired people. It is a natural language for them and is not a burden" (Ohki, 1). Sign language is the natural language for those who need to use it, similar to spoken word is for us. For this reason it is quite desired in an interface, as a verbal interface is for those who can speak. Not only is it possible to use sign language to interface with a computer, but also more importantly a sign language translator can be used for communication with people who are unable to understand sign language. Speech recognition software could also be used in the future to translate spoken language to a sign language display.

Gesture and face recognition can be used for those who have limited mobility. Software has been created that detects blinking, regardless of the head angle, it is possible to sense head motion and facial expressions that could be used to interface with a computer by a disabled person (Reilly, 20). Although for the average person gesture recognition would not be successful as an interface in itself, it will be quite useful in combination with other interface methods and also serves a purpose in the development of artificial intelligence.

 

5. USING BRAIN SIGNALS

Using speech and gestures to input into a computer is quite advanced and extraordinarily useful, but what if a computer could be interfaced purely using brain waves. Using the brain, as an input device, would eliminate the difficulties for everyone in using a computer, assuming the user has a fully functional brain. For those who are unable to communicate or move due to some sort of disability, "the only possible method of communication would be to use electrical signals produced by the brain as a switching device for computer interaction" (Polikoff, 178). With a system like this many disabled would be able to use their mind and finally be able to compete and communicate to the outside world, not only that but it could make their life more livable with virtual worlds to enter into. Not only would brain waves serve useful for the disabled, but also for the future of the fully interactive interface.

So far by using the scalp recorded mu rhythm Polikoff, Bunnel, and Borkowski, have been able to achieve one dimensional cursor movement (178). Although this form of input is not yet accurate it is rapidly improving. By sensing the mu signal, the P300 detection system has been 50% accurate but this has been improved somewhat over user practice. (Polikoff, 180)

This field is relatively new and is showing signs of improvement. As of now we have only discovered one electrical signal that can be used for this type of interface. As more signals are found we will be able to control systems in more ways. The possibilities of using brain signals, for an interface, are almost endless.

 

6. THE INTERFACE USING A HUMAN FACE

So far we have discussed how we can make a user interface more usable, especially for those with disabilities. Although we may be able to create a computer that takes only your brain to operate, it will still needs to have a friendly output. It has long been the desire of computer users to have a more humanistic interface (Walker, 85). Although there are many simple ways to make an aesthetic and friendly interface, for instance the color and the placement, I wish to discuss the research done on how a human face can be used in the interface. After all, "The most human-like interface … would embody human characteristics in an agent with human form, of which the human face is one of the most compelling components." (Walker, 85). With the increase in computer technology, speech synthesis and facial animation has been made possible and we are now quite able to make an interface that appears like a human face.

The research done by Walker, Sproull, and Subramani shows us that using a human face in the interface will actually have an effect on the users. The results of the experiment showed that there were more negative results with the human face than text. The stern expression had the most negative results and the neutral was in the middle with text producing the best average results. (Walker, 88)

This is a table displaying the results of the comparison of text with the human face interface. The values are on a scale of 1-10 with 1.0 being the most negative unless otherwise noted. This table was taken from Walker, Sproull, and Subreamani page 88.

	Text (n=15)	Neutral Face (n=15)	Stern Face (n=12)
Attitudes to answering questions
Were questions clear?	8.1	7.1	6.2
Were you comfortable?	8.7	8.2	8.3
Similar to face-to-face?	2.2	1.9	1.9
Want to continue?	4.6	3.4	3.0
Involvement
Time spent (minutes)	24.1	26.8	35.0
Requests to repeat (freq)	N/A	6.5	8.0
Missing answers (freq)	3.6	3.3	2.7
Invalid Answers (freq)	8.1	4.9	3.2
Unsolicited comments (words)	57.1	24.9	57.2
Open-ended questions (words)	65.6	57.8	114.9
Final comments (words)	30.0	37.9	68.3

Although we see that the response time, the question clarity, the comfortableness, and the desire to continue were a lot lower in the human face interface we also found that errors and missing answers were less common with the stern face and second with the neutral face. We also see that the requests to repeat were the highest with the stern face (Walker, 88).

These results show us that the human face seems to have an effect on the user for evaluating answers and the more expression on the face led to the better evaluation of answers. Although the user did not necessarily like the stern face, it produced the answers with fewer mistakes. The lower number of mistakes is due to our natural respect for a stern face and our feeling of being evaluated, causes us to pay more attention and produce a better result, regardless our contentment with the system. (Walker, 89-90)

 

7. CONCLUSION

There are a large number of emerging technologies especially designed for those with disabilities that will soon be incorporated into standard user interfaces. It is easy to realize that many of these technologies will eventually surpass the use of only disabled but also by the general computer user population. Using speech recognition in an interface is well desired by the computer user population. Gesture recognition is still a fairly new technology for users to understand the capabilities, but it will also be highly desired in the future when it is properly marketed. I think possibly the most fascinating technology would be the use of brain signals to communicate with a computer. This is a technology that is just starting to have credibility, but may eventually be a standard just like using a mouse.

These many devices allow for a disabled user to be able to use a computer and compete in the business world. It also gives a temporary escape from their disability where they can slip into a virtual world. Computers provide a means for all to be treated equally and they give everyone the possibility to thrive in our society. It is important to continue research in these areas to assist the disabled, but also to advance the way we use computers.

We also find that the display of a human face effects the way we interact with a computer. Personable software will keep the users happy and productive. It is important in an interface to mimic the persons view of reality, this way they do not feel isolated from the computer, but instead we feel we are interacting with the computer, as one would with a human.

 

8. REFERENCES