Thursday, February 22, 2007

How many megapixels do our eyes have?

BBC Focus Magazine has a regular Q&A section, which, unfortunately, they do not publish on the Internet. The March 2007 edition has a reader's question - "How many megapixels do our eyes have?"

Their answer approaches the question from multiple angles. I won't quote the entire response here, but some of the interesting "factoids" are:
  • comparing the number of sensing elements - the eye has 5 million cones (the color receptors) and 100 million rods (the monochrome contrast receptors) which give a human the equivalent of two 105MP (MegaPixel) video cameras (because we have two eyes).
  • comparing spatial resolution over the eye's field of view - we have the equivalent of 576MP.
(Source: BBC Focus Magazine)

Sunday, February 11, 2007

Hey, someone changed my system beep!

I could have just as easily titled this post "Much Ado about Nothing", but it wouldn't have been as informative. PC Magazine does a good job dissecting the various weaknesses of the "flaw" in Microsoft Vista security which would allow an audio-based attack.

Ultrathin folded telephoto lens


Engineers at the University of California San Diego have created a folded lens by machining reflective surfaces into a optical crystal. this technology has the potential to shrink lens assemblies, which would be beneficial to lots of devices, such as cameras and night vision gear. However, this is still early stage research and hurdles remain - such as its inherent narrow depth of field. That being said, there are ways around that problem that might work for selected applications where thickness is more important than length (where I imagine that additional lenses with different focal lengths could be positioned on a rotating disk or similar arrangement to be employed as needs changed) or post-processing power consumption.

The work has been published in Applied Optics (February 1st, 2007)

(Hat tip and image source: PhysOrg)

Security holograms relatively easy to copy

The Daily Irrelevant has a post on the rise in counterfeiting of security holograms, which are found on everything from credit cards to whiskey bottles these days. Reportedly, the cost of the equipment to make a copy is in the range of only $2,500 USD, which, given the illegal profits to be made, is insignificant. Interesting reading.

Saturday, February 10, 2007

Facial composite study results

According to a recently released study, facial composite systems used by police do not produce very good likenesses - not even good enough for people to correctly identify celebrities. The systems they studied were ones the police use to "build" (or composite) a face by having the eye witness select from a large collection of eyes, ears, noses, etc. EurekAlert reports that
According to authors, these poor results are not deficiencies in the software per se but instead a mismatch between how we remember faces and how composites are produced. "Numerous lines of evidence converge on the view that faces are generally processed, stored and retrieved at a holistic level rather than at the level of individual facial features." Ultimately the psychological process of remembering faces may include more complex representations such as multidimensional similarity to other faces or relative sizes and distances of features and so on that are not readily retrieved by memory nor utilized by facial composite software.
The authors go on to recommend whole face, or "looks like", methods instead of the existing "parts" based methods (e.g. eyes and ears) as a way to get better results. "Looks like" systems have produced good results in previous studies, as I recall, so maybe they are on to something.

The study was published in Current Directions in Psychological Science (February 2007 issue).

Saturday, February 03, 2007

Noise map of Great Britain

The Telegraph (a center-right newspaper in the UK) has published a noise map of Great Britain that was compiled from traffic noise levels by Deepak Prasher, professor of audiology at University College London. I can not recall ever seeing such a map, but it is fascinating to not only note the absolute levels on particular places (especially the sites where you need ear plugs to get a good night's sleep) but also to compare different areas.

Training your mind and dyslexia

One of my favorite radio programs is the excellent Science Friday, which is broadcast on NPR (National Public Radio) in the USA. A recent edition talked about the plasticity of the human brain and included some interesting statements about research into treatments for dyslexia. I had not heard of the studies linking dyslexia to hearing difficulties - not vision processing difficulties as one might assume. You can find an MP3 download at the link above.


Clarification: I should have said "linking dyslexia to auditory processing difficulties" or something similar, not "hearing difficulties", which was too ambiguous. I should have also said "not only vision processing difficulties", as some theories of the roots of dyslexia still focus on vision processing (for instance, the so called magnocellular theory).

The root causes of dyslexia are not fully understood and serious research is ongoing. I suggest the following Nature article (Franck Ramus, 2001) for an overview.

I would also like to say "thank you" to the readers who pointed out my misstatement and the ensuing discussions in the comments. If you are just now coming across this post, I recommend reading the comments to catch up.

Military gunshot detectors

Strategy Page has a brief overview of the current state of gunshot detectors used by the US military. It talks about both all-acoustic and acoustic-infrared systems. Enjoy.

Thursday, February 01, 2007

Image blurring of sensitive information

First, the attention grabbing bit - using traditional digital image blurring techniques may not be effective in all cases. The title of the article from which this post derives is "Why blurring sensitive information is a bad idea".

Why? With statistical techniques, it may be possible to extract some of the original image details were. Now for the complicating details. First, it wasn't shown with a real case - it was simulated. Second, although one could probably deduce this from the article, traditional digital blurring is in all likelihood very effective against information with high spatial frequencies (i.e. images with lots of changes, like pictures of faces not taken at a close distance). The problem MIGHT occur with blurring data with mostly low frequency content, like large sized numbers, such as on an image of the check used in the article.

Nonetheless, it is a very interesting piece of work. I look forward to someone showing this work against "real data".

Hubble is hobbled

This is widely reported, but I still think it is important enough not to pass without mentioning - the Hubble Space Telescope has been hit by another fault, this time from an electricalshort. To summarize the the most probable outcome, it will probably lose two of its three"camera channels", leaving it with only the Solar Blind Channel (which is sensitive to ultravioletwavelengths). To read more, here is the New Scientist Space article.