Just on the right side of Impossible (part 2)

In Just on the right side of Impossible, I mentioned several ideas that I think could be ideas for frighteningly ambitious start-ups. Three items came up in the news this week related to some of these ideas, here they are:


* Communicating with Animals : Is a Dolphin a Person? There is a slow recognition that animals are getting to a status we never thought they had. We know they have language, we empirically catch some of the signs but there is no "thing" that permits us to communicate efficiently with pets and other animals. Once we have that "thing", the question as to whether a Dolphin a Person? becomes far more frightening.


* Real Telepresence: Skype recruiting Xbox developers for 'next generation services  The issue at hand is the total disruption of the travel industry. I expressed some of my thoughts on the subject earlier see Great Thoughts Friday: The Energy and Time of Being There. and Kinect Teleconferencing with Real-time 3D Capture and 3D Display

On a technical note one wonders if the XBox/Kinect is really what is needed. One wonders if compressive imaging could help reduce the Kinect bulk.

* New Types of Sensors: 

A UWB motion detector and a Phone camera that can "see" radiation:

* Xandem's security sensors can see through walls. The Tomographic Motion Detection (TMD) described here. Here is the video presentation:



  Here is a demosntration





Of interest are the attendant publications that led to this product:

• N. Patwari and J. Wilson, "People-Sensing Spatial Characteristics of RF Sensor Networks", arXiv:0911.1972v1 [cs.NI], Nov, 2009.
• J. Wilson and N. Patwari, "Through-Wall Motion Tracking Using Variance-Based Radio Tomography Networks", arXiv:0909.5417v2 [cs.OH], Oct, 2009.
• J. Wilson and N. Patwari, Radio Tomographic Imaging With Wireless Networks Tech Report, 16 Sep 2008
• J. Wilson, Device-Free Localizaton Using Received Signal Strength Measurements in Wireless Networks, University of Utah Department of Electrical and Computer Engineering, Aug. 2010.

Remember, with Cable, we mentioned the ability to extract radiation data from a webcam. Well, I noticed an android App named GammaPixLite trying to do the same.. I note these interesting notes (emphasis mine):

Notes: 

A 20-minute initialization is required before you use the application for the first time. We know this step slows you down, but it really is necessary to get the best results. Please perform this step in a place you know is likely to be free of excessive radioactivity.

For best results make sure no light is getting into the camera when you run the GammaPix Lite app. Putting the phone in your pocket or covering it with a book works well.

A reading takes about 5 minutes if there's no danger. Dangerous levels will be reported sooner.

The GammaPix App will not work on some phone models (e.g. Motorola Atrix) because the camera optimization at low light levels gives "bright" pictures.

Not all phone models have been calibrated. The reported range will be large for devices without calibrations. Try out the app anyway. We'll rush calibrations for the most popular phones running the GammaPix Lite App.

This free Lite version requires a network connection to run and expires June 30, 2012. A full version with more features is planned before then, but we wanted to share what we've already built!

The Robust PCA approach we take with Cable is precisely that you can take a video of something that is not moving and get a radiation reading. Here the app forbids any amount of light to get on the CMOS chip. Let us note that this is not the same as a gamma camera aka a coded aperture camera. In either the webcam or the android application, we do not put anything in front of the sensing system to localize the source. Therefore, both the webcam and the android app are providing a radiation measurement with no information about the source of that radiation. 

Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.