02 March 2009

Gesture Recognition – Breaking the Barrier between us and our tools/toys

Historically, humans have required textual/graphical user (keyboards, mice) and hepatic/tactile (joysticks, data gloves VR helmets) interfacing systems to interact and communicate with computers. The latter group of devices sense body movements and “feed” information to software based on position. Conversely, users can receive information in return through the aforementioned hepatic channels in the form felt sensation such as a vibration (amalgamated from http://www.webopedia.com/TERM/H/haptic.html and http://www.webopedia.com/TERM/I/interface.html). These input/output tools have typically chained users to their machines, requiring them to be within arms length of the devices in order to interact with them. Gesture recognition technology is striving to end this bondage, by not only improving interaction between user and hardware but also by opening doors for individuals who are physically challenged and have difficulties with the current user interfaces.

Gesture-recognition systems identify human gestures (typically head, arm, hand, and leg movements in addition to torso position) via the use of a camera (and in some cases motion sensors or ultrasound emissions and infrared lights) that read the movements of the human body and conveys the information to a computer that uses the gestures as input to control devices or applications. Though a majority of the systems can only interpret broad gestures (problematic…more later) some are capable of decoding facial and speech expressions (i.e., lip reading), and eye movements(amalgamated from http://www.webopedia.com/TERM/G/gesture_recognition.html and http://ieeexplore.ieee.org/ielx5/2/29695/01350721.pdf?arnumber=1350721.

The technology’s proponents tout a variety of potential uses for the technology including but not limed to:

• Surgical simulations
• Improving security and surveillance
• Military applications
• Interpretation of sign language to spoken word
• Replacement of traditional input devices- keyboards, mice, joysticks etc.
   o Virtual Keyboard for PDAs; http://www.ideinc.com/canesta.html
   o Head- and eye movement-based cursor and mouse interface 
       technolology, - users can work with applications by moving
      cursors  with head movements and clicking the mouse with eye
      blinks instead of double-clicking, users can double-blink while 
      looking at an icon or file name. http://www.cybernet.com/products/navigaze.html
   o Automobile controls drivers would change, for example, the 
      temperature or sound-system volume by manoeuvring their hand in   
     various ways over a designated area. This could increase safety by
     eliminating  drivers’ current need to take their eyes off the road to
     search for controls  http://www.cc.gatech.edu/ccg/projects/gestpan/ 
Gesture recognition technology has faced and still faces major hurdles from intrusive hepatic devices (as not all systems have switched to camera, sensors etc.) to its resource intensity resulting in the interference with other system functions and slow image processing/lag which is undesirable when playing a dynamic video game. Furthermore, many products can’t read motions accurately or optimally when faced with changing or excessive lighting or when the user is situated near or in front of “busy” backgrounds. An additional key issue lies with the technology’s inability to interpret which intended gesture a series of motions actually represents; this is compounded by the lack of a common gesture language, specifying the way users should make gestures to make sure they are easily recognized. http://ieeexplore.ieee.org/ielx5/2/29695/01350721.pdf?arnumber=1350721.


Research, continues to improve the field for example, improvements to algorithms resulting in faster, more robust, and more accurate systems. Additionally, hardware and processing costs have decreased considerably, whilst processor speeds, sensor accuracy and camera recognition have increased. When hepatic devices are required technology has made them smaller, wearable and less intrusive for example gesture recognition sensors are being placed in rings. One of the more innovative improvements in the field comes from GestureTek who used heuristics (a branch of artificial intelligence) to achieve more “robust, accurate, and quicker tracking of gestures” by turning multiple cameras onto a subject; thus allowing enhanced data analysis of gestures in three dimensions http://ieeexplore.ieee.org/ielx5/2/29695/01350721.pdf?arnumber=1350721.

On February 6th GestureTek the pioneer, patent-holder and world-leader in camera-enabled gesture-recognition technology for presentation and entertainment systems invited our class to their Toronto office where we were given the opportunity to learn about and interact with cutting edge gesture technology.
The company's multi-patented video gesture control technology (VGC) lets users control multi-media content, access information, manipulate special effects, even immerse themselves in an interactive 3D virtual world – simply by moving their hands or body delivering Wii-like gesture-control without the need to wear, hold or touch anything http://www.gesturetek.com/aboutus/corporatebackground.php
The company features seven products (http://www.gesturetek.com/pdfs/Gesturetek_fact_sheet.pdf) used across a variety of industries including:
• museums, science centers, amusement parks, aquariums, zoos, visitor centres
• television production, game development
• trade shows, real estate presentation centers, corporate showrooms, boardrooms, virtual 
  videoconferencing
• advertising - retail locations, digital signage, interactive multi-media displays, kiosks
• bars & nightclubs and other public spaces such as airports and stadiums
• surface computing solutions, mobile
• healthcare sector http://www.gesturetek.com/marketuses/industryuses.php

Though GestureTek got its start in the healthcare sector as a result of Vincent John Vincent (one of the co-founder’s) background in physical rehabilitation and his desire to develop an engaging solution that could hasten patient recover; today GestureTek heavily focuses on advertising and gaming.

I personally am interested in the work they have been doing with museums and other cultural centres http://www.gesturetek.com/marketuses/museums_sciencecenters.php as my background is in education and technology. I view education as being broader then just what is taught in the classroom or formal system, museums and other tourism and cultural facilities have an enormous opportunity at their fingertips to engage their audiences, if they will step outside of their “traditional boxes.” Additionally I would like to see GestureTek expand their gaming division to focus on serious and educational game development; the learning potential due to the experiential nature of their technology is limitless, the simulations that could potentially be created could revolutionize not just corporate level and post secondary training but also if made affordable enough k-12 truly enabling teachers to become facilitators rather then gatekeepers of information/knowledge.
A relatively extensive internet search (over an hour of Googling) confirmed what GestureTek told the class about having little competition in the field. Reactrix a company who licensed GestureTek’s technology and then began competing against them selling interactive advertising displays went bankrupt in December 2008 http://www.technologyreview.com/communications/14508/?a=f. According to Venture Beat http://venturebeat.com/tag/invde-shaw-co/ several other companies pursuing the same goals have remained afloat. The major example is Catchyoo, a Japanese company that projects massive interactive ads on floors, walls and surfaces. Israeli 3DV Systems are poised to pick up where their fallen rival left off. There are also lots of companies, such as Danoo, that are putting up flat-panel displays at places such as Starbucks. The videos and other ads on the screens can entertain people while they’re waiting in line. Extreme Reality http://www.xtr3d.com/ XTR developed proprietary real-time high-resolution software that analyzes 3D human motions using one simple web cam, with out any additional accessories while Oblong Industries, Inc an LA-based company is using gesture recognition software in conjunction with gloves, cameras, and 3D sensors that track position to replace the mouse by following a user's hand and finger positions and orientations which are then parsed and interpreted by Oblong's proprietary gesture recognition engine for interpretation http://elianealhadeff.blogspot.com/2008/04/gesturetek-and-oblong-serious-gaming.html

No comments: