$59 Smart Camera Module (Pixy)

By Ben Wood, Published Aug 27, 2013, 12:00am EDT (Info+)

Beset by lawsuits, the video analytic industry has seen little innovation recently. Now, a new offering from Carnegie Mellon researchers offers a super low cost smart camera module, raising over $270,000 on Kickstarter. We talked to Richard LaGrand, founder of Charmed and CMU partner, about the technology and its possible applications. 

Background

******* **** **** *** ****** ********* to **** ********* ** ********** *** to *** ******** ******. 

“**** * *** ** ****** * remember * *** **** ******** *** had ** *** ** ******* *** $25,000 ** *** * ******** ****** system. **** *** ********** *** ***** rate *** ********* **** ***** **** you *** *********** ***** ** ****. What **** *** ** ** ****** basic. ** *** **** ** ****** a ******** ****** **** * ******** station ** ***** ***** * ****, it *** ** **** **** ** thing."

*** **** ****** **** *** ** provide *** ********** ******** ****** ********** for *********, *********** *** ********* ***** robotics. *** ******* ******* ** **** and ** ** *** ***** *******. It ** ** ***** **** ******** than *** ***** ******* *** ***** a ******* ** *** ***** (********* $59 ** ***********, *** *** ***** backers *** ** *** $**).

***** ***** ***** **** ***** (*******: it ****** **** *** ******** ******):

Possible ************

**** *** ***** ** ********* **** other *******:

“***** *** * *** ** ***** ideas *** ****** **** *** **** readers ** ** **** *******. ** has ******** ** ** ***** ** **** as * ******** ******, *** ****** are ****** ****** ** **** **** work **** ******** ********** ** **'* hard ** **** **** *** *********** is,” ** ****. “*** *** ** 10 ****** **** ** ** *** do **** ********* ... ** **’** committed ** ********* **** **** *** road (***** ********-***** ********* [**** ** ****** *********]).” 

***** ** ** **** *** ******** applications? ********. *** *****'* ******* *** not ********* ** ***** ************. ** said *** **** ***** *** ** develop *** ********** *** **** ** open ****** ** *** ****** ****** apply **** ** *** ***** ** applications.  

“** *** **** ** ******** * low **** ****** **** ****, *** be **** ** *** ****** *** of *** ******, *** ** ** cheaply, ****’* **** * **** *******. I ***** **** ****** *** ********** to *** **** **** ** **** solved **** **** ********** **** ** ********* **** *** use ** ******** ********** ****'* ******* been ******* ***.”

How ** *****

**** *** ****** **** *** **********, using ** ********* ** ****** ****** color ********** ** ********* ****. ** stores ***** ********** ** ***** ****** and **** **** ** ******* ****** images. 

** **** *** ****** *******, *** can ****** ****** *** ****** *** findings ***** ** ******** ****** ***** to ***** ****** ** ***** ************. This ***** **** ** ********* ******* colors:

*******, *** ********* ** ********** ******* it **** *** **** **** ** object ** * **** ** * box ** * ***, ***. **** ****** detect ****** ** *******. **** ******* ****** are *****, ******** ****** ** ******* (e.g., *** = ****, **** = box).

“*** **** ***** **** ** *** button *** **** **** *** ****** in ***** ** ** *** ** ‘learns’ *** ****** ***’** ********** **,” LeGrand ****. **** *** ***** ********** for ***** ********* ****** *** *** detect ******* ** ******* ****** ***** signatures **************.

Capability *** ***********

********* *****

 “*****’* *** ****** * *** ********. It *** **** ** ** **** how **** ** *** ***** ** view ** ********,” **** *******. “********* on *** ***** ****** ** *** lens, ** ***** ** * ****** small ******.” *** ******* ** **** on *** *********** *** * *.** mm **** *** *** * ** degree ***** ** ****. 

**** ********* *********** ****** ** **** a ******* *****. ** **** ** object ***** ** ** ***** *** pixels, **** *** ****** *** ***** information.

*** *****

******* ******** **** “*** **** ****** of ***** ** *** * **** image. *** ****** ** * ******* inch ******. ** **’* *** ****, but *** *** ******, *** **’* pretty *********. **** ** **** ** though ** **** *** *** ***** it. ** *** **** **** ******* to * *********** *** ***********, ** might ****** ****’* ******* ** ******** that ******, *** ***** **’* **** to **-*****, ** ***** **** **** issue.”

**** ******

*** ********** **** **** **** ****** that **** ****** ***** ********** ** browns *** ***** *** **** ********* to ******, ******* ****. “**’* **** possible **** ***** ******* ***** ** some *** ** *****, *** **’* best ** *** ** ****** **** has * ****** ***** *********.”

Potential ************ ************ 

***** **** ***** * **** **** and ****** ***** *** *******, ***** its ******* ************ (**** ****** *** certain ****** *** ** **** *** conditions), ** ** ********* ** ******* mainstream ************ *** *****. 

Tech *****:

  • *********:***  *******, *** ***, **** ****
  • ***** ******: ********** ******, */*", ********
  • **** *****-**-****: ** ******* **********, ** degrees ********
  • **** ****: ******** *** (******* ********* types *********)
  • ***** ***********: ************
  • ***** *****: *** ***** (**) ** unregulated ***** (** ** ***)
  • ***: **** *****
  • *****: ** *****
  • ********* **** *******: **** ******, ***, I2C, ***, *******, ******
  • **********: *.*" * *.**" * *.*"

Comments (4)

That's awesome! I can imagine in a short time the technology that will come from this.

Ok - that is awesome!

Cool article, thanks IPVM. Here's my comment on it. While the technology it shows is very exciting, it almost appears to be more suited for machine vision than it does security. My reasoning behind this is that because Pixy detects colors and hues, and combine that with the fact that lighting conditions are generally dynamic in surveillance environments, this would generate a lot of false positives and/or negatives.

One application might be retail, in a PoS type application where Pixy is validating PoS data with expected color data from the packaging passing over the scanner. Just a thought.


But the $59 cost point is very intriguing indeed.

I can see this product being used inside of toys such as dolls or animals. With the capability of following learned objects, the toy can go after the ball, bone, mouse, or whatever accessory that comes with the toy. And, it would be an infinite use accessorized toy since how Pixy can learn other objects. As for example, if you have a dog toy, with a Pixy camera built in, it could play with a ball, and if you lose the ball by accident, all you have to do is replace it with any other one in your house. You just need to learn it into the Pixy system. Plus, being able to watch what Pixy watches on you pc can make for great home videos of the toy playing with your child or family. This product can be used for a wide range of ideas and purposes because of how great and affordable it is right now.

Login to read this IPVM report.
Why do I need to log in?
IPVM conducts reporting, tutorials and software funded by subscriber's payments enabling us to offer the most independent, accurate and in-depth information.
Loading Related Reports