The Development of Speech and Facial Recognition

by Modis on April 17, 2012

Many games and applications for the iPhone and Android platforms toy around with facial recognition software, which currently can only display and detect 2D models. Not only is it limited, it’s sometimes inaccurate due to reflections and lighting on faces. Speech recognition has become a blind person’s or slow typist’s dream. With advancements, computers can actually learn new vocabularies and be taught to respond to different voice patterns with proper training.

Photo Source: familymwr w/CCLicense

Historical Milestones

A rudimentaryform of face recognition software has been around since the ’60s, though it had to be directed to certain features of a figure before it could identify the form as a face. [10] Similarly, speechrecognition began one word at a time – literally. A word was spoken followed by a pause each time to allow for processing. [11]

By the ’70s, scientists were capable of identifying faces that could capitalize 21 unique features. Unfortunately each of these faces had to be manually constructed and placed into the system. [10] By this time, pauses between words became obsolete in speech recognition, and research progressively enhanced the user experience until 1980 when it began making its way onto telephone hotlines. [11] As of 1991, modern facial recognition software began to develop, though at the time, it was extremely limited by environmental factors. [10]

Photo Source: planetc1 w/CCLicense

Practical and Controversial

The earliest practical use of facial recognition was 10 years later for controversial scanning of SuperBowl attendees in Tampa, Fla., in early 2001. Further experimentation using automated cameras was done in the Ybor City district to catch criminals and miscreants littering the streets at night. By summer of that same year, Virginia Beach had been granted $150,000 by the state of Virginia to invest in facial-recognition-equipped cameras to possibly identify missing children and criminals. [1

Two years after the cameras were installed, Tampa took them down due to their ineffectual and shoddy recognition software. Virgina Beach still maintains their cameras today, despite their failure to catch any miscreants since their inception in 2002. [1] in 2003, the speech recognition market reported revenues in excess of $677 million and from 2004-2006 surged to over $1 billion in reported revenues. [11,15]

At the airport in Boston, a study was conducted to verify whether or not the facial recognition cameras actually worked. Of 249 tests, the software in question only recognized the participants 153 times. Subsequently the program was scrapped following this evidence supporting the lack of a consistently accurate technology. [1]

Photo Source: ukhomeoffice w/CCLicense

REAL ID and Terrorism

Due to the terrorist attacks on the World Trade Center towers in 2001, a proposed law was passed in 2005, called REALID, which makes it possible to implant a chip inside any driver’s licenses that were printed as recently as 2008. This chip supports facial recognition to determine if the person with the license in their possession is actually the owner of that identity. [2]

Speech recognition has found its calling in both the government and medical fields. Ingovernment, speech recognition software is used to reduce paperwork and increase productivity. Some software guarantees 99% accuracy straight out of the box while being able to process a little under 100 words per minute. [12, 13]

Only in the last 6 years has facial recognition made a serious comeback via United States immigrationservices. A digital scan of fingerprints and the face is required for any incoming foreign travelers. The information is then compared to the biometric information on their passport for verification. Should someone be unwilling to go through these procedures, they will typically be rejected from entering the country. [3]

Photo Source: daniel_iverson w/CCLicense

Social Networking

In addition to the government’s use of the software, Google has implemented facial recognition as a feature into Picasa, a photo uploader and editor. After uploading an image and associating the various faces to names, Picasa will search your desktop for familiar facial features and add them to the corresponding albums. [4]

Today, facial recognition is used to automatically tag photographs on Facebook and various other controversialways. [5] Many of the future integrations and advancements in this field are based upon social platforms using your facial features to log into their sites rather than actual passwords. Various devices have begun substituting fingerprints in place of traditional log-ins. There have even been discussions of Android and iOS eventually supporting features that utilize the facial recognition technology as well as fingerprints. [6]

Pioneer has developed an Android app called Whodoo that uses voice recognition to post on various social networking sites as well as to read messages received by various services. It also offers maps and directions via speech. [14]

Medical Field

Previously unfathomable practical uses for 3Dimaging has resulted in numerous breakthroughs in the medical field. For extreme cases in which a person’s face has been disfigured, scans were taken to ascertain the extent of facial tissue damage and determine whether or not a person is capable of receiving a transplant. The images produced from this ensure that the donor’s face matches anatomically with the recipient. This technology is then used to customize the skin to match the patient’s facial features. [7]

Speech recognition has made a remarkable impact upon the medicalfield in various ways. First, telephone services can now communicate, ask questions, and resolve a caller’s problem using a specialized computer. Additionally, reports and other paperwork can be typed using your voice and be seamlessly stored electronically. [11]


It’s difficult to predict where facial recognition will go from here due to the unpredictability and moralebattles being waged against future endeavors. [8] Facial and voice recognition technologies may have practical uses in modern society but the inaccuracies the facial recognition technologies currently allow make them bad candidates for security solutions today or in the near future. [9]

















Related Posts

Leave a Comment

Previous post:

Next post: