Sign Language Translation Glove

Published:
Updated:
The SignAloud Glove is capable of translating American Sign Language signs into text and audio.

Background


Recently, there have been developments in aids that can enable members of the hearing community to communicate effectively with the Deaf community. In order to understand this topic and its significance here are a few statistics. ASL (American Sign Language) is the fourth most “spoken” language in the US. One in eight people in the United States (12.5 percent, or 30 million) aged 12 years or older have hearing loss in both ears, based on standard hearing examinations. About 2 to 3 out of every 1,000 children in the United States are born with a detectable level of hearing loss in one or both ears.


Although the idea of a device that can interpret and translate languages in real time is nothing new, one particular invention has stood out among the rest. Please understand: there have been numerous attempts at developing an invention with the ability to translate American Sign Language, but one in particular, SignAloud, has garnered more attention than its predecessors. The SignAloud translates ASL letters, numbers, and signs in real time. This may seem like nothing new compared to the other inventions of its kind, but the SignAloud glove is lightweight and practical if needed for everyday use. Interpreters can oftentimes be expensive to hire and many businesses, even places of employment, do not provide interpretation services to the Deaf/deaf and hard of hearing. The SignAloud provides users with an easier design and more cost effective alternative that features sign to text and sign to speech options.


About the Data


The SignAloud glove collects the following data: pressure, hand, arm, finger, and wrist position, movement, frequency, and time. All these things are measured by the gloves sensors. These pieces of data are used to interpret what has been signed by going through a database and searching for the correct “values”.


Example of Use

   

John Doe is going to visit a new physician next week. He knows that he won’t able to afford an interpreter and that the clinic will not provide one. He sits and thinks for a moment. A few moments later he remembers hearing about the SignAloud glove. John orders one and uses it at his doctor’s appointment. When the doctor begins asking John questions, he connects his glove to his tablet and starts to sign his responses. The glove takes the data from his hand position, movement, and pressures through the numerous sensors that adorn the glove. From there, the data goes into the stored database and converts the values into spoken words.


Limitations and Considerations*

   

In its current state, the SignAloud glove seems that it will only be able to translate basic signs. In order for the glove to be truly effective, they would need to make sure that the sensors used are very aware of slight changes in pressure and position. Some signs have a slight visual difference but have very different meanings. The signs for the letter F and the number nine look very similar, the only difference being whether the thumb and index finger are rounded or flat and if the fingers are together or apart. 


Numbers one through five are signed with the palm facing the signer and the remaining numbers through nine are signed with the palm facing out. If this piece of knowledge isn't known, then the glove can mistake a W for a six or a two for a V. To ensure accuracy, sensors may need to be attached to the arm and in between the fingers in order to capture a wider variety of signs. 


In addition to what was mentioned above, a lot of time must be spent programming signs into the database and converting them into numerical values. ASL has over 10,000 signs! This number doesn't reflect the various dialects that are found in the United States. ASL is a dynamic language so there will always be new signs added and ones that become obsolete. Even if the changes are every few years, the database of every glove would have to be updated. 


There are different ways to say the same things so that must be taken into account as well. Just like English, certain phrases in sign language can be said in numerous ways. For example, you could sign: "I'm fine.", "I'm doing okay.", or "I'm well". To many, these sentences have the same meaning but have a slight difference. The glove would also need to be calibrated for dominant and non-dominant hands because some signs require the use of both hands.


Because the grammar and syntax in ASL is conveyed through facial expressions and body language, sensors would be needed to be placed on the face and chest to truly reflect how ASL is "spoken". The only way to circumvent this problem would be to require users to employ interrogatives whenever asking questions. On a similar note, developers would need to know the approximation of a person's body to effectively recognize signs. 


For example, the sensors would need to be able to differentiate between male and female signs. When signing family words like mother, father, brother, sister, and cousin, the signs are the same. The sign for mother and father are the same, as are brother and sister. The only difference is where the sign is placed. Most family signs follow a male-female pattern. All male signs are signed at the top half of the face and all female signs are signed at the bottom half of the face.


*In the interview with Navid Azodi and Thomas Pryor ( Found Here: YouTube Presentation ), the two inventors never mention these aspects. It could be that they didn't want to go to in depth with their summary and have made progress towards fixing these problems or they are not aware of them at all.


2
1,850 Views

Comments (0)

Have a question about something in this article? You can receive help directly from the article author. Sign up for a free trial to get started.