August 1993 - cypherpunks-legacy

Real time OCR
by The Phantom 12 Aug '93

12 Aug '93

--------------- cypherpunks-list #4087 (54 lines) --------------- Date: Thu Aug 12 08:40:56 1993 From: wcs(a)anchor.ho.att.com (Bill_Stewart(HOY002)1305) <wcs/daemon> Subject: Re: [uk.transport] Speed Camera with OCR > > The ability to read a car's plate & decypher the number in real time > > without any human intervention must take a bit of work. > There's a lot of automated-highway research going on; most of it has > been oriented towards radio+smartcards, which are easy to make accurate, > and while they're not totally cheap, you can palm off much of the cost > on the car-owner and use cheap equipment in the tollbooth. > Bill Stewart If anyone can dig out the April issue of 'Advanced Imaging' (I seem to have given mine away), there was an article about imaging license plates for use at border crossings. The bugs looked like they were worked out and the system was ready for action. I don't remember where it was going to be put into effect, but.. mt Matt Thomlinson Say no to the Wiretap Chip! University of Washington, Seattle, Washington. Internet: phantom(a)u.washington.edu phone: (206) 548-9804 PGP 2.2 key available via email or finger phantom(a)hardy.u.washington.edu

1 0

WIRED #4
by Sandy 12 Aug '93

12 Aug '93

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ SANDY SANDFORT Reply to: ssandfort(a)attmail.com . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Punksters, Paul Ferguson asks if the new WIRED is out yet. The answer is "yes and no." Complimentary copies were given away at a couple of recent conferences. Also, subscription copies are going out now or in a few days. Newsstand copies should be available in three weeks or so. I would appreciate any Cypherpunk thoughts, opinions, questions or raves on my article, "Intelligent Island?" which appears in this issue. It and William Gibson's "Disneyland with the Death Penalty" are the lead articles in WIRED #4. Hope you like it. S a n d y >>>>>> Please send e-mail to: ssandfort(a)attmail.com <<<<<< ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1 0

CryptoStacker
by Kent Hastings 12 Aug '93

12 Aug '93

CryptoStacker#000# Whatever happened to CryptoStacker(s) research? Please voluntarily comply with my request for a status report, or I'll padlock your business, throw you in jail, arrest all your friends, and firebomb your compound. It is your choice. Kent - <jkhastings(a)aol.com>#000#

3 2

Re: Secure voice software issues
by norm＠netcom.com 12 Aug '93

12 Aug '93

Eric Blossom <eb(a)srlr14.sr.hp.com> says: > I have seen estimates that a straight forward implementation requires > about 13.5 million Mulitply+Accumulates / second. Most of the time is > burned up using a brute force search for the best excitation vector to > use. There is a fixed 512 entry code book, and a dynamic code book > with 256 entries (it may be 128). Each code book entry is an > excitation vector that is 60 samples long. Therefore, to evalute each one, > you have to run a 60 element vector through a 10 pole filter to get > the predicted output, then compute some measure of error. This > requires an additional difference operation that is implemented as > some kind of "perception weighting filter" (I don't remember the > details). I have been reading the PowerPC 601 manual (MPC601, The Macs of early 1994). It is dangerous to believe performance figures. They give you the world in one chapter and then take it back here and there in bits and pieces. Here is what I see however. Simple single precision floating point operations can issue one per cycle. The book mentions several floating point ops that take more than one clock in a pipeline stage. They don't mention floating multiply-add. I think one can issue each clock. I-unit instructions can issue in the same clock as floating point ops. If you do the block trick used to multiply matrices then one load is required per multiply add. All this leads to the optimistic estimate that the 50MHz machine can sustain nearly 50 fmadd's per microsecond on a 50MHz chip. Inner products are much like matrix multiply which is a benchmark where the RS/6000 (The MPC601's father) achieved nearly one fmadd per clock, and that was double precision! 128 excitation vectors each of 60 single precision loats fit in the on chip cache, but it is tight. There may be enough margin here for it to work with no special DSP. I'll be in Yosemite for a few days so I won't be able to respond immediately to comments.

3 2

under pressure from University Computing svcs,
by hfinney＠shell.portal.com 12 Aug '93

12 Aug '93

From: The Phantom <phantom(a)u.washington.edu> > I'm going to have to shut down the phantom remailer. > > Seems someone complained about a remailed message. I got a copy of the > message, and it didn't seem to bad to me, but I've gotta go with what UCS > says. > > As of today, the phantom remailer (phantom@mead) is hereby out of service. > > Matt It's too bad that this remailer is being shut down. I posted a few days ago about this problem, and someone asked if there had been any specific examples of shutdowns. Here is one; I know Matt's is not the first. If it would help, Matt, I could provide you with code to act as a "second- tier" remailer, one which would only forward to one of the other Cypherpunks remailers. No person would receive messages from your remailer, hence there would be no complaints to your administration. Your remailer would just be an extra "entry port" into the Cypherpunks remailer system. I don't know whether the terms of the shutdown of your remailer would allow you to experiment with what is arguably a different piece of code, one which would not lead to complaints. Hal Finney hfinney(a)shell.portal.com

1 0

Re: Chaos harnessed for encryption / Fluctuations and Or
by fnerd＠smds.com 12 Aug '93

12 Aug '93

gtoal(a)gtoal.com sez- > That was smoke and mirrors by people who understand chaos better than > they understand encryption. All it boils down to is a synchronised > pair of (not very good) PRNGs. > ... > Forget it. It's a dead end. What it is is PRNGs that can synchronize without publishing their states. If this could be done with strong PRNGs, you'd have something. -fnerd quote me

1 0

RE:patent on remote cash transactions?
by nobody＠alumni.cco.caltech.edu 12 Aug '93

12 Aug '93

THE MOUSE THAT ROARED. A small company has received a big patent covering financial transactions from home computers. While industry giants have bickered over who's on first in home banking, shopping, entertainment, etc., Online Resources & Communications Corporation of Fairfax, Va., has received what appears to be an enormously broad patent covering financial transactions made from home computers. "The patented payment process potentially applies to a wide range of nonbanking services, from pay-per-view television to stock trading done from home terminals." (New York Times 8/9/93 C2)

1 0

Re: Secure voice software issues
by jim＠tadpole.com 12 Aug '93

12 Aug '93

V.42 adds about 200ms to my slip link. I'd rather do without it, thanks.

1 0

Re: [uk.transport] Speed Camera with OCR
by wcs＠anchor.ho.att.com 12 Aug '93

12 Aug '93

Rick Tait, R.Tait(a)bnr.co.uk, reports Marcus Reynold's note that there's a new > > speed trap in Kent. Apparently the radar clocks your speed, a TV camera > > records your number plate & then a real time OCR system picks out the > > actual number & flashes it up on a digital display a short distance > > further up the road. [...] > > The ability to read a car's plate & decypher the number in real time > > without any human intervention must take a bit of work. There's a lot of automated-highway research going on; most of it has been oriented towards radio+smartcards, which are easy to make accurate, and while they're not totally cheap, you can palm off much of the cost on the car-owner and use cheap equipment in the tollbooth. License-plate reading in real time is tough, if you're trying to get both speed and accuracy from a video image of a moving car; our neural-net folks were looking into it about three years ago, and it seemed like they'd be pretty good at it if they could get a bit more funding:-) ; they've mainly been concentrating on envelope-readers for Post Offices and railroad-car identifiers. Embarrassing speeders is a special case, because the accuracy doesn't really have to be 100% - you catch the cars on video, so if you decide to prosecute anybody, you can have a human read the picture and verify that you had the right license plate number. Meanwhile, back on the road, if your computer misidentifies the speeding driver half the time, he may assume the speed trap was nicking the other fast car next to him, and you've scared a lot of other drivers into slowing down because it *looks* like you've could have nicked them; so you've accomplished the safety objective. It's scary stuff, because the technology is mostly there, and a government that wanted to pay for development could get it done by a number of research companies. If it's too expensive for widespread use now, computer horsepower keeps getting cheaper every year, especially when what you need is DSP-crunchers and gate-arrays for neural nets, rather than general-purpose systems. Meanwhile, the smartcard people are mostly not using digicash, they're using no-privacy systems and appealing to the convenience for the commercial driver, whose company probably doesn't object. Dave Mandl comments: > Hmmm...most drivers I've known would be thrilled to get off with just > "embarrassment." In fact, they'd probably be proud to have it broadcast > to everyone else on the road (except the lawmen, naturally) that they'd > just been clocked at 110 MPH or whatever. A number of years ago, Ohio tried a system that displayed your speed (but not your license plate) as you went by - they quickly discovered that they had to cut off the display at something like 80 or 85 MPH to prevent just that effect (though adding the license plate number would cut down on the exhibitionism a bit.) Bill Stewart

1 0

Re: [uk.transport] Speed Camera with OCR
by dmandl＠lehman.com 12 Aug '93

12 Aug '93

> From: R.Tait(a)bnr.co.uk > I have just seen an item on the the BBC South East News about a new > speed trap in Kent. Apparently the radar clocks your speed, a TV camera > records your number plate & then a real time OCR system picks out the > actual number & flashes it up on a digital display a short distance > further up the road. The idea is to embarrass drivers by letting them > & others see their speed & registration. I don't know what type of > equipment they are using, but it sounds like a really neat trick. > The ability to read a car's plate & decypher the number in real time > without any human intervention must take a bit of work. Has anybody > heard how reliable this system really is ? Hmmm...most drivers I've known would be thrilled to get off with just "embarrassment." In fact, they'd probably be proud to have it broadcast to everyone else on the road (except the lawmen, naturally) that they'd just been clocked at 110 MPH or whatever. But somehow, I doubt that this data is used only for embarrassment, don't you? --Dave.

2 1