This is much the same concept as recently presented at ICDSC-2009 in the paper "Abnormal Motion Detection in a Real-time Smart Camera System", based on real-time video analytics and artificial neural networks to autonomously build a model of what is considered normal behavior, subsequently flagging outliers as possible events that require further (human) scrutiny. Thus it acts as an intelligent self-learning classifier/filter that greatly reduces the information stream per camera for human operators.
Indeed, OCR still has a long way to go to make it work
reliably with live camera views, as is stressed also
at the Mobile OCR for the Blind
page. It at least gives some proof of concept for others
to improve upon. The open software interface makes it
easy to try other command-line driven recognition engines.
Anyone who knows of some useful recognition engine to
plug in there?
By the way, complementary to the remote sighted guide
idea, there is also mobile camera phone
software that gives blind people direct access to their
visual environment, while including a talking color
identifier. It appears to work on the Nokia 6600 and
various other camera phones.
That's right, and the software for seeing with sound is available as
The vOICe Learning Edition
It can be complemented with verbal feedback
from automatic OCR, face and object recognition
once reliable vision-based recognition engines
become available, but at present it still seems
beyond the state of the art in computer vision
to do this in typical real-life environments.
Related experimental technology for the blind is also available for free elsewhere ("The vOICe"):
Mobile OCR for the blind
includes speech recognition and speech synthesis
support. Currently the proof-of-concept demonstrator uses the GOCR OCR engine, but other (object?) recognition engines can be easily added.
Stereo vision for the blind
Slashdot Mirror
Smart Fingertip Camera to Help Blind People http://www.infoniac.com/hi-tech/smart-fingertip-camera-to-help-blind-people.html At first glance this looks like prior art
I'll gladly cooperate with Sergey Brin on "Seeing With Your Ears" http://www.seeingwithsound.com/ to "hear" an otherwise inaccessible environment
Good points, John. There is still a long way to go for visual prostheses to become usable, convenient and safe options even for totally blind people.
It's called "Talk to Me" by Flaviu Negrean, and it currently supports speech-to-speech translation from English to Spanish, French, German and Italian. Works well on my Android 1.6 phone. http://www.androlib.com/android.application.info-fidogames-apps-talktome-zwnB.aspx
This is much the same concept as recently presented at ICDSC-2009 in the paper "Abnormal Motion Detection in a Real-time Smart Camera System", based on real-time video analytics and artificial neural networks to autonomously build a model of what is considered normal behavior, subsequently flagging outliers as possible events that require further (human) scrutiny. Thus it acts as an intelligent self-learning classifier/filter that greatly reduces the information stream per camera for human operators.
Already available at http://weocr.ocrgrid.org/
Indeed, OCR still has a long way to go to make it work reliably with live camera views, as is stressed also at the Mobile OCR for the Blind page. It at least gives some proof of concept for others to improve upon. The open software interface makes it easy to try other command-line driven recognition engines. Anyone who knows of some useful recognition engine to plug in there? By the way, complementary to the remote sighted guide idea, there is also mobile camera phone software that gives blind people direct access to their visual environment, while including a talking color identifier. It appears to work on the Nokia 6600 and various other camera phones.
That's right, and the software for seeing with sound is available as The vOICe Learning Edition It can be complemented with verbal feedback from automatic OCR, face and object recognition once reliable vision-based recognition engines become available, but at present it still seems beyond the state of the art in computer vision to do this in typical real-life environments.
Related experimental technology for the blind is also available for free elsewhere ("The vOICe"): Mobile OCR for the blind includes speech recognition and speech synthesis support. Currently the proof-of-concept demonstrator uses the GOCR OCR engine, but other (object?) recognition engines can be easily added. Stereo vision for the blind