Slashdot Mirror
Digital Cameras vs Scanners for OCR?
ttennebkram asks: "With 6 and 8 Megapixel cameras on the market, some now with Wifi built in, it might be more convenient to shoot pictures of your bills and papers with a camera than fussing with the scanner. By the numbers, it would seem feasible. 300dpi for an 8.5"x11" sheet of paper works out to about 8 megapixels; 300 dpi is usually what OCR vendors suggest. I imagine for high volume good results you'd want to maybe mount the camera on a tripod arm over your desk. Heck, I was thinking of a glass desk and maybe one camera below and one above, and maybe a foot pedal to trigger the cameras (and I suppose a flash and high F-stop would help as well). If I could quickly 'snap' all the junk paper I have and electronically file it, maybe OCR the images at night in batch while I'm asleep, and then maybe get rid of all that paper once and for all. Using a traditional cheap scanner just takes too long. So has anybody tried this? I realize that camera optics are different than scanner optics, so maybe it's not just a question of raw pixel counts. Any thoughts?"
...the aspect ratio and even lighting are your enemies. It's almost impossible to shoot a bill or a check stub dead on, at close rage, without fish-eye'ing, and without getting in your own shadow. Sure, you might have a little white linnen box that you use to take your eBay photos, but, seriously, this is a job for a scanner.
I haven't used cameras that would really be called "good" but it was really hard to get the text perfectly in focus and readable even with the camera on a stand. Also, lighting was kinda a pain but it sounds like you have that under control.
Is it just me or is it not going to upgrade to Vista in here?
What you want is a scanner with a sheet feeder and a GOOD one at that. They're not that expensive anymore, since there are lots of cheap machines which have a feeder anyway due to them having a fax function. This alone will go faster than manually swapping the papers and shooting with a camera.
8 of 13 people found this answer helpful. Did you?
The problems I had were (a) getting the book flat, and (b) getting the lighting right. With flash, you end up with a ring of brightness and by OCR software got very confused, as the grey newsprint outside the flash's ring was being handled as black.
If I were a whizz with Photoshop/GIMP/etc, I suppose I could have done some sort of correction to the picture, but...
I've heard how Kinko's have book scanners that will copy and bind a book for you - perhaps they also have a scanning to CD/DVD service? Would that be cheaper for you?
The scanner is in the top, instead of the base. Flip the lid over, and you can scan without lifting the lid every time. Could probably scan sheets almost as fast as you could with a camera setup. And far cheaper.
Some people report not being able to get good scans with it, but I've had no problems.
I pay (almost) all of my bills with my online bank and just get e-bills. Then there is no paper -> OCR -> file issue. It's just website -> file. Saves you a big investment too!
While cartoonists used cameras suspended from tables to make animiations for years, I don't really think this is what you want. A better solution would be to buy a scanner with an Automatic Document Feeder. These aren't particularly expensive and are usually worth it for other reasons. The most cost effective ones are the Laser Multifunction devices. They are basically laser printers with a USB port so they can act as a printer or scanner as well. Just make sure you get one with an ADF; it's worth it. Most also throw in fax functionality as well. My Brother MFC-7420 has lasted me over a year (still going strong) and two toner cartridges and can scan a batch of papers at 300 dpi in no time flat. While it prints B&W, it scans in color at a max of 9600 dpi, which is good enough even for pictures. It only cost me $250 and was totally worth it.
Google uses cameras to digitize their books, so it's feasible. I would shoot from 3-4 feet away with the lights at a 45-degree angle from the glass on either side.
dom
If it's just for keeping a record of bills and other junk, why even bother with OCR? As long as you can read the results, just snap away.
I was conned by an old man in a cloak. It turns out those *were* the droids I was looking for.
C'mon, your work doesn't have a scanner/photocopier/printer with a feeder? I take my paperwork into the office once a quarter or so, feed the lot through the scanner in the print room, and email the output to myself at home. If you're one of the rare cases who'd feel bad about this, you could always offset the expense by not using their water cooler or coffee for a week :)
Get a scanner
If a job's not worth doing, it's not worth doing right.
http://www.sentex.net/~mwandel/tech/scanner.html :)
I quite often take photos of receipts and forms rather than walk to the other room to boot up the PC with the scanner. The trick for me is to force the focus to 1 metre and take the photo from about a meter away while using the optical zoom to make the paper fill in the field of view.
I admit that the quality isn't quite as good as the scanner, but it's a heck of a lot more convenient and it's good enough for many uses.
FWIW, macro mode doesn't really work well on my old digicam, plus, I'm not sure it makes sense to use it in this case.
It really sounds like you haven't actually tried it. I've got a Powershot 450, I think they just upped the versions so they're past that now, I imagine. Anyways, it's 5~mp, and I used it all the time in the library to take pictures of books so I could accurately quote them...much much easier than copying long quotes word for word, and then just look at the picture and re-type the text.
My suggestion is to not take things so far digital with your process. Paper doesn't take up so much space that you can't hold onto digital and physical copies:
1) Gather your big stack of "documents I probably don't need but if I do need, I will really really need"
2) Take digital images of them all, either with a scanner or a digital camera. Choose a method, try it a few times, check the output to make sure it works and is legible, and then go with it.
3) File the digital image on your computer, and file the paper in a system with the same setup and title. Even if you put all the papers in a an office-style cardboard box and stick them all in the back of the closet, the attic, whatever, do it in the same system as best you can.
The point here is that you use the computer files if you ever need the information, and if they fail, you've still got the paper in the back of the closet in your basement. If you keep the papers organized, you can store them in such a way that after a few years, you'll know that an entire box is safe to shred and recycle.
The key to any system like this is being able to trust that you have what you need when you need it. You can't ensure digital or physical documents 100%, but with both, you can feel pretty safe (store a copy of the digitals at a relative's house or some such, somewhere significantly off-site, or on an FTP somewhere or somesuch, or course). But having the files isn't the only part of a reliable system: you also have to be able to find them.
I have some experience doing what you're trying to do. I've even done this type of work in professional labs with serious pro equipment (it was my job). It's a huge pain in the butt.
I'm currently digitizing my collection of old tabloid punk magazines from the 1970s. I had to use a digital camera because flatbed scanners that do 11x17 or larger are extremely expensive, they're like $3000 or more. So I did some experiments with my consumer-grade 5 megapixel digital camera. The results were adequate, barely (and I have an art degree in Photography, this stuff is easy for me, YMMV). I've currently suspended my project until I can afford a higher rez digital camera, mostly because 5Mp is barely enough to capture the little 6 point type that is used in large sections of the magazines. But let me tell you more generally what I've learned.
First off, you'll need a copy stand. This is a fairly standard photo accessory, but a good copy stand is fairly expensive. You need something that is easily adjustable, so you can raise and lower the camera to get the document to fill the frame, without using too much zooming. The copy stand keeps the camera parallel to the target at all distances. It is important to have quick adjustability in height, rather than zooming. You'd be much better off using a "prime lens" rather than a zoom, as zooms tend to have barrel and keystone distortion.
Secondly, you need lights. If you only want to copy written documents (or B&W magazines like me) you can use cheap spotlights. If you want to do color, you need much better lighting, something with a fixed color temperature, or a flash system. Spotlights are really hot, and when I work in my small office, it gets intolerably hot when I spend about an hour photographing. For better, more repeatable results, you'd be better off getting a flash system. BUT...
Here is the sticking point. You need something to keep the documents flat. That means placing them under a sheet of glass. So you are going to get reflections from the lights, and flash is high intensity lighting which makes it even more difficult to control reflections. The usual method is to put polarizing filters over the lights and the lens, to cancel out the reflections. This is a rather complex method, and a LOW END professional copystand with polarized lighting will set you back about $2500.
OK, so what I did is I adapted my old disused photo enlarger. It was a huge monster for 4x5 negatives, I took off the enlarger head, and used a Bogen photo clamp with a ball-head joint attached to the motorized arm that goes up and down. It does a fairly good job as an improvised copy stand, but it is pretty cramped, the baseboard is only designed to make max 20x24 prints. Also it is a HUGE pain in the ass getting the camera leveled with the baseboard, I use a bubble level. Then I attached a cheap set of tungsten photofloods to the wings of the enlarger, so the light hits the baseboard at a 45 degree angle, to reduce glare. Note that it is best to point each light at the far side of the document, so the light paths cross each other. This gives the light a little distance to fan out and eliminate hot spots. I don't put my documents under glass, they're newspaper pages, so I flatten them for several weeks (!!!) under weights, then if there's a little curl, I use weights (like heavy metal rulers) at the edges, or hold the edges down with post-it notes. That eliminates the need for a glass plate to hold them down, and I don't have to deal with reflections. However, it takes a LOT of time and effort to get the documents positioned and flattened correctly, it is not a quick process.
I use a Canon camera, so I use the Canon Camera Remote to my laptop to preview and take the shot. Even with the lights and some fill flash, I can end up with exposures of 1 or 2 seconds, so I can use a narrow f-stop. This shouldn't be necessary for a flat object, which requires no depth of field, but I find that the lens is sharper stopped down. It takes quite a bit of fiddling to get the optimal
I'd agree with you, if I didn't have the impression that the original poster wants to scan things that are not of a 'regular' printed size. Most consumer grade ADFs are designed to only handle A4 or letter.
Insert
so get yourself an A-size scanner and just scan each page in two parts?
Or if there aren't too many grayscales that you'd trash,
just run it all through a photocopier to shrink to 8.5x11 and scan that?
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
ScanRprovide a way for you to experiment. basically, they take an image, ostensibly from your cell phone, but could be from any digital camera and converts it to a searchable .pdf file.
.pdf.
Depending on the quality of teh camera, they sugegst conference room whiteboards; Diagrams, notes, flow charts
Their website has samples of what you'd expect, but basically you're capturing the image and sending it to scanR. They do the conversion and send it back as a
I'm curious: Is parallax and other image distortion a significant problem when you take pictures of a paper vs. scanning it?
http://outcampaign.org/
Is there any software available that can roughly OCR a document and store keywords or snippets of text in metadata or an index? This would solve both the problem of both keeping an accurate record and searchable content.
What formats allow an easy mix of image and text data (without formatting)?
Python coder | PyQt Applications | Writer
So to clarify... You want to trade the hassle of:
1) lift a lid
2) stick a paper in a well-defined corner
3) press a button
for the hassle of:
1) align a camera on a tripod, including angle as well as position
2) align a paper with no guide
3) adjust the lighting so that you get an even tone
4) make sure you didn't accidentally move the camera, the tripod, or bump the desk
5) step on a foot pedal that you jury-rigged to make take a picture
OR
5) Push a button on a camera that you can't afford to move even a hair.
6) Use image software to continue adjusting the photo so that the OCR will read it properly
7) Hope you did everything right the first time.
I think I'd pick door number 1.
"If you make people think they're thinking, they'll love you; But if you really make them think, they'll hate you." - DM
I have a four year-old Canon Powershot G2 that has been indispensible in the digitising of my documents. Given adequate lighting, all you need is to line up the document in the view-finder and take the photo. Autofocus is usually adequate, but if you just can't seem to get a clear shot (certain things will prove problematic), manual focus is your next best feature to utilise. If you're doing James Bond-type work and are in a hurry, then you'll often end up with blurry images that won't be useful in OCR. Given that I already own a digital camera, I will never invest money in a scanner. If anything, I'll buy a better camera when I can find one with all the features I want. Hope this helps in some strange way :)
A lot of people seem to be critisizing the idea, but there are some uses for it.
I have a multifunction scanner/printer/copy/fax that cost around $100 when it was purchased with a computer a year or two ago. Its great for scanning in receipts for work expense claims, and having soft copies of important paperwork. I used to hand-hold a digital camera, with receipts and papers on a well-lit, flat surface, and photograph in macro mode. Now I'll be going back to that method as I've moved into a very small place where space and power are limited. I don't have room for the big bulky device, nor a filing cabinet for storage of the originals.
Long story short, taking the pictures by hand, without a tripod means that yes the image of the papers will often be distorted (taken from an angle) or have shadows in them. But the point is you can read all of the lettering and 9 times out of 10 throw away the original. I would never bother with OCR, just keep the image files well-organized and backed up.
Your Brain + EEG + LEGO Robots = Brainstorms
I was thinking of a glass desk and maybe one camera below and one above, and maybe a foot pedal to trigger the cameras
Boy, you're right! Who'd want to fuss with a scanner!?
I'm against picketing, but I don't know how to show it.
"What formats allow an easy mix of image and text data (without formatting)?"
http://www.lizardtech.com/
DjVu is what some libraries use and there's some free software out their.
BTW there are legal size scanners out their. Shop around, plus don't forget to check manufacturers websites for refurbished and discontinued models.
I find that most bill providers have an option to receive your bills electronically, keeping them either in their "safe" (ie, website) or to receive them in e-mail. This is true for credit cards, banks, major utilities; the main exception being the city-run water and trash company.
I routinely scan pages of old books (and other documents) for my Web site, from old books. I use an Epson Expression 10000 XL, which, as someone else noted, isn't cheap, but it does A3/11x17/tabloid at 2800dpi. At 400dpi grayscale it can scan a regular page in a few seconds.
I've also used a Casio Exilim camera to photograph pages.
The way that it's done for archival purposes is to have a mount that holds a book and also holds a medium-format camera about four feet away. To get good resolution for OCR you'll need something that's about an 11 megapixel camera or more, for a full page at (say) 7x10 inches of actual text. Hugin and ptstitcher and friends, the panorama tools, include software to correct for lens distortion. Phase One sells a camera mount (in Canada you can get it from Vistek, together with their 40 megapixel back end for a medium-format camera. Or you could make a suitable mount yourself. The trick is that it holds the book open half-way (or less, using mirrors) so that you don't get as much page distortion. Holding the book and the camera rock steady is absolutely necessary if you are photographing text.
For small items like a cheque (say), use a flatbed scanner, and scann at 400dpi grayscale. Project Gutenberg's guidelines are outdated (they use 300dpi black and white as I recall) and don't get such good results. If you go much higher than 400dpi, the OCR software starts having tantrums at you and the quality may actually degrade.
The best OCR software on the market today as far as I can tell is Abbyy Finereader. I tried several, and found this had, for example, at least two orders of magnitudes fewer errors than the GNU OCR package. You should expect errors, though, especially in digits.
Frankly I'd go with a scanner just because they're designed for this application, and you have less hassle. Transferring images from the camera to the computer twenty minutes after taking the photo means you need to keep a separate log of where each photo came from, or you'll muddle them up. I save images with filenames like Ball-Sussex/086-Pevensey-Castle.png so that the page number is in the filename. And the image quality with even a low-end scanner is much higher than you can get in practice with a camera without an elaborate set-up, and reliably better, comes out every time regardless of lighting, camera settings, wobbly hands, etc.
Having said all that, I do photograph pages sometimes to make manual transcriptions. Afterwards I do careful proof-reading against the original.
Liam
Live barefoot!
free engravings/woodcuts
It's almost impossible to shoot a bill or a check stub dead on, at close rage, without fish-eye'ing, and without getting in your own shadow.
If his purpose is simply not to file paper a scanner is not required. A $200 Cannon from Walmart is all you need if you don't worry about OCR. You move the camera back and use the zoom and it works. I take 1600x1200 pictures of my classnotes and the results are perfectly legible. A good desk light saves your batteries by eliminating the need to flash. You move it to the side to avoid your shadow. The result might be fisheyed, but so what it's just your phone bill. Being able to OCR and text search would be like icing on the cake. The real prize is keeping your records without stuffing file cabinets with unimportant junk.
Friends don't help friends install M$ junk.
These units scan both sides of the document in the same pass, at between 4 and 30 monochrome sheets per minute depending on resolution (up to 600 DPI) and model. They can also do color (and grayscale), but that tends to be slower and the files MUCH bigger. The Xerox is faster, but more persnickety; the Fujitsu is unflappable and optimized for one-step desktop use. With either one, you can just drop a stack of pages, checks, receipts, or whatever into the sheet feeder, press a button, and PDF files appear on your disk. This is VASTLY more convenient than feeding, or photographing, one sheet at a time, and the scanners are small--footprint is not much bigger than a sheet of paper.
Speed isn't as important as you might think: unless you're scanning huge stacks of paper, the time spent fussing with the paper itself tends to dominate the process, and it's easy to do something else productive while the sheets are whirring through the auto-feeder. I've used them for everything from manuals to business cards, and I think they're the ideal solution to getting rid of all that paper. The software is idiosyncratic and not extremely stable (I occasionally have to reboot the machine that the Xerox unit is connected to), but it gets the job done.
I'm currently digitizing my collection of old tabloid punk magazines from the 1970s.
That's hard to do but it's not what's required. Snapping legible pictures of a phone bill is not hard if all you want to do is get rid of your paper. Getting OCR is harder, but still not as difficult as making museum grade preservation of artwork.
Be sure to post the results on line some time after the copyright expires. In 2070, they will probably read like Elizabethan English but at long last the public domain will be served thanks to your efforts.
Friends don't help friends install M$ junk.
Ray Kurzweil as tweaked a digital camera to do just what you describe, for the purpose of providing a portable text reader for blind people. They are for sale now, for about $3500 each.
If you're doing low volume work, using a camera may be fine. However, once you start getting into the higher volume work, you want a scanner with a document feeder. Also, non-sheetfeed scanners are generaly cheaper than 6-8 megapixel cameras, so I'm not sure why one would use it. Especially since OCR really only needs B&W and any camera of that quality is going to be color only. A black and white scan takes a lot less time than a color.
Back to the sheet feeds, I've worked with the Fujitsu fi-5220C series scanners before. The only time I've ever had a problem with the document feeder was when I forgot to remove a staple or paperclip. It's also quite fast, 3 seconds tops for a BW scan of an 8.5x14 and that includes the time to transfer the scaned image to the PC. I'd challenge you to find a camera that could keep that up for long, especially as you would have to manually change the pages yourself.
Fly me to the moon Let me sing among those stars Let me see what spring is like On jupiter and mars
With that mentality, why don't you hope they buy out a grocery store, or a gas station?
Following World War II, a lot of photography businesses would photograph the discharge and service records of US servicemen and women, so they could present a copy of their records when applying for various benefits offered to veterans. My grandmother did this (the photography) after the war. She was a veteran as well - she worked for US Navy Intelligence working on Japanese military cyphers. She also analysed aerial photographs of enemy positions, but I digress.
Unfortunatly, my grandmother passed away, so I cannot get any information from her on how the photographs were taken, but one of my family members is working on a biography of my grandmother, and she might have the information.
Zagreus sits inside your head, Zagreus lives among the dead, Zagreus sees you in your bed and eats you in your sleep.
It took a bit of digging to find it again, but I ran across this in some huge chain of events that is way off topic, but the result is one translator's approach to using a digital camera to scan his manga collection without unbinding. The foot petal idea is what tipped off my memory.
1 .php
http://www.mrdummy.net/mangatranslation/tutorial0
I am, and always will be, an idiot. Karma: Coma (mostly effected by
Hi this is Chris from scanR.
v ery_pi.html
It's nice to see discussion on this topic. We think cameras are a much better consumer platform for capturing physical information and making it accessible digitally. There are roughly 600 million cameras + camera phones shipping each year and they are easy to use. So if you don't have a scanner, and you certainly don't have one in your pocket, using a camera is a great substitute. Of course, scanR makes it better, including OCR keywording for search.
Read this post on what makes a good camera for scanning: http://blog.scanr.com/scanr_blog/2006/04/making_e
I have done document capture using a scanner and a digital camera.
It falls into three main steps:
- 1. Acquire the image via camera or scanner
- 2. Fix image quality
- 3. (optional) OCR the image
I fix the image quality (i.e., reduce colors, fix the spotlight effect of a digital camera's flash, etc) by:
1. Stretch the image's contrast using ImageMagick (similar to automatic level adjustment)
convert.exe input_image_0.png -equalize image_1.png
2. Reduce the number of shades of white. (changes all pixels within 10% of white to be white)
convert.exe image_1.png -white-threshold 90% image_2.png
3. Reduce the number number of shades of black. (changes all pixels within 10% of black to be black)
convert.exe image_2.png -black-threshold 10% image_3.png
4. Adaptive threshold the image. Important that you set the window (WidthxHeight) to be about the height of two lines of text and about 4 characters wide. A smaller window runs faster but can cause noise to be put inbetween two adjacent lines of text.
convert.exe image_3.png -lat 20x20 image_4.png
5. (optional) OCR image_4.png
Other options:
- Threshold by a percentage
--> 50% threshold using
convert x.png -monochrome out.png
--> User definable threshold amount (e.g., 80%) using
convert x.png -threshold 80% out.png
Lastly, you can chain together commands so that the steps can be simply written as one line:
convert input_image_0.png -equalize -white-threshold 90% -black-threshold 10% -lat 20x20 output_image.png
Noise removal can be optionally done before any other image processing
Noise removal methods:
- NL filter for 'edge enhancement' in gimp (Filters -> Enhance -> NL Filter) (Use settings, Edge Enhancement, Alpha(0.60) Radius (1.0)). This is similar to the pnmnlfilt.exe in Netpbm - found at netpbm.sourceforge.net
- Despeckle (not always helpful)
- Dust and scratches in Photoshop
- Smart blur in PhotoShop
- Unsharp mask - It does a good job smoothing background noise but does not do a very good job at edge enhancement (NL Filter does a better job).
ImageMagick is open source and can be had here http://www.imagemagick.org/
Netpbm can be had here http://netpbm.sourceforge.net/
I did this in two projects:
- Scanning a large number of oversized newspapers (10 inches by 12.5 inches) (serveral hundered pages)
- Scanning in 8.5 inch x 11 inch tinted color pages from books (about 500 pages, messy because pages had a color background)
The multiple scan for an oversize page takes too much time because you have to scan the page face down on the scanner.
This requires you to:
1. lift up the paper,
2. flip to the next page,
3. flip over,
4. place on scanner (align it)
5. scan (suggest 300 dpi or higher black and white (monochrome) scan)
6. align page for second scan
7. scan (suggest 300 dpi or higher black and white (monochrome) scan)
8. (later) put both scans into a single image file, or, much more time consuming, join the two scans into a single seamless image
This takes about 1 minute per page given a 10 second per scan scanning time.
A digital camera would take about 10 seconds per page without any post processing, page joining in an image editor, etc.
The main OCR issue with a camera is that you need to approach the quality of 250+ DPI scanner image. You need to take a camera snapshot with a dpi higher than 250 to get near scanner quality because camera will distort the image, add noise, over sharpen, etc. I've had good results taking images that work out to 350 dpi with a digital camera.
They had USB scanners at the local Goodwill for about six bucks. Six bucks! 'Course, you had to dig through a box of wall warts, and at least one of them had a mangled drive belt, but if you're willing to bring a laptop, it's hella cheap.
Laws do not persuade just because they threaten. --Seneca
It'll be too big. I have a standalone USB-powered scanner that's 1" thick. (It's now retired because my printer came with an integrated scanner.) It was great when I was in college, because I could stick it in my backpack and take it to the library.
It sounds like you want to perform scanning in a batch job. Perhaps an off-the-shelf solution is better, even if it's slow? (You'll be asleep, at work, ect.) Do the old HP ScanJets allow for batch scanning?
No, I will not work for your startup
1 Don't use a tripod, use a document photo stand.
Think of an overhead projector with the camera where the mirror is for vertical adjustment.
2 Have a guide for the paper, not that hard.
3 Lighting is an important one, but as long as it's even the type of light doesn't really matter if you set your white balance correctly.
4 If it is a rigid setup doesn't really matter
5 Use the camera control software on the computer, you don't need to really use a camera.
6 Save the file and run the OCR software.
I use a similar setup to take photos of test parts at work, works nicely.
i belive for in print books even google are cutting off the spines and shoving them through a sheet feeder, far less labor that way.
yes if you are archiving rare books there isn't much choice but for most applications sheet feeding or flatbed is fine (yes flatbed without sheet feeding is laborious but i'm not convinced theese "planetery scanners" are any less so)
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
OCR seems to work with Goatse.
_______
|=(0)=|
| | | |
It is a good idea to scan the receipts you get from the cashregisters when buying something with a warrenty.
It's often in odd formats, making it hard to punch holes in an storing it in a binder.
But more important: If the receipt is printed on cheap thermo-sensitive paper, the print fades away.
What good is a 2 year warranty, if the receipt is just a blank piece of paper after 3 months.
Leif