Team
COMPONENT ONE
Part One
The problem our team has selected is voice identification. Voice identification is not easily accessable to everyone therefore we are doing research to help make voice identification devices affordable.
Part Two
If we can make voice identification devices more accessable and affordable, we can totally eliminate the concept of keys. In doing this, we can make a normal person's life, easier and more efficient.
Part Three
We find that this problem is not just one that occurs is certain locations, it is a problem that seems to be happening all over the United States and in other coutries as well.
Part Four
This problem can effect our lives by engaging a new risk-free source of security. There are no identical voices, therefore we can insure the fact that these devices are dependable and secure. If we can integrate these devices into our daily lives, we can allow ourselves to be more carefree and only worry about more important things.
Part Five
Because voice identification devices aren't inegrated into our daily lives, the chances of getting locked outside your car, losing your keys, forgetting pin numbers and combinations are more likely to occur. If we had voice identification devices, we wouldn't have to worry about keys, memorizing combinations, and we wouldnt have to call a lock smith because your stuck outside of your car.
COMPONENT TWO
Part One
Biometrics was first used by the Egyptians. Of course they didn't have the technology we have but their way of keeping a record was clever and barely had any flaws. They took note of people by listing who their parents were, significant features, etc.
Part Two
Two scientists/engineers that made major contributions to this technical application is William C. Dersh and Dr. E.A Quade, who demonstrated the SHOEBOX (voice recognition arthimetic calculator) experiment.
Part Three
Our technical application, biometrics directly impacts our problem. Our problem is how to make voice recognition systems more affordable and more user friendly. We think that improving biometrics technology, making it smaller and using less expensive materials, should make it more popular and spreading it’s use from businesses to everyday homes.
Part Four
We have found in our research that although voice recognition systems are more dependable than ordinary locks, it also has it's major disadvantages. Our voices though distinctive may not offer enough information for speech-based identification. It is also a very sensitive system, it requires that it's environment be free from loud noise and that the system be fitted for the users physical state (example:their height). Though it may have it's bad side, it is very popular with people especially in business. It is 1 of the 4 highest accepted biometrics systems out there.
| Biometrics | Universality | Uniqueness | Permanence | Collectability | Performance | Acceptability | Circumvention |
| Face | high | low | medium | high | low | high | low |
| Signature | low | low | low | medium | low | high | low |
| Voice Print | medium | low | low | medium | low | high | low |
| F. Thermagram | high | high | low | high | medium | high | high |
Signature-Impossible to copy dynamics
Voice Print-High Accuracy
F. Thermalgraph-Superior to face recognition, non-intrusive, plastic surgery can't fool it
COMPONENT THREE
Part One
To get a very high quality recording needed to be able to verify a person's voice efficiently you would need equipment in an everyday business is not usual. And because of the possibility of outside noise while recording and verification you need a sound-proof environment and the recording equipment would not be effective enough. An employee of the business might be sick and their vgoice might chasnge for that time. The computer would read his voice and it would be different from the stored information. That would cause denied access or confuse it with another employee's stored model of their voice. So to make it more effective, you could. Combine it with another way to identify a person such as fingerprints. Since people don't have the same exact voice and same exact fingers, they would be a combination of already good identity verification systems.
Part Two
By combing fingerprinting and speech recognition, it would be more effective in verifying a person's identity. Something such as a cold or any other illnesses would not have as much of a threat to confusing identities or not giving access.
Part Three
For this piece, we contacted Dr. Anil K. Jain, Director of the Pattern Recognition and Image Processing Lab at MSU. We contacted him via e-mail and over the telephone.
Part Four
According to Dr. Jain, our proposition for a small lock (i.e. a door lock) functioning via voice recognition instead of a lock or combination is highly feasable, and not too complicated. Microcphones have certainly gotten small enought to be used for our purposes, and all that would be needed would be a small amount of memory inside the lock itself and a chip that'll compare the stored voice sample to the input voice sample. That said, we could also reduce the size by having the calculations and comparasins be done from a computer instead of within the lock itself. Even a desktop PC could suffice, you would only need to add the required comminications equipment within the lock in order to achieve this. On the topic of pricing, something of this nature could be avaliable for less than $100, Dr. Jain estimates around $30-$40. If we could create this highly feasable project, and get it out into the mass market, it could be possible to integrate voice recognition systems into our everyday lives and not have it seen as a costly service that only corperations or companies can afford to implement into things such as car or home locks.
Components
