Poor OCR results MG6250

Hello,
I used my new MG6250 and its MP Navigator EX for character recognition of a text from a printed book.
Unfortunately, the results are not particularly satisfactory, lots of mistakes (about 30) per page are shown.
I sent the scanned image of the original text to a free online ocr and received an almost perfect result.
Is the MP Navigator EX ocr really that bad or might someting be wrong with my settings?
Thanks
Michael
My system:
Windows 7 Home Premium (64 Bit)
Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz

Hi Michael!
Welcome to The Canon Forums, and thank you for your inquiry! We appreciate your participation, however we need to let you know that your product seems to be a model that is not supported by our team here at Canon USA. The Canon Forum is hosted and moderated within the United States by Canon USA. We are only able to provide support for Canon products manufactured for the US market. If you live outside the United States, please click here and select your country or region for your support needs. Feel free to discuss Canon products sold outside of the United States, but please be aware that you will not receive support directly from Canon USA.

Similar Messages

In the evaluation mode im nt able to view my ocr result for more than 2 seconds

im doing my final year project on image processing using labview 2010(evaluation). im using a microsoft webcam(VX-700). my objective is character recognition using vision. im able to capture an image and store the same. the software also recognizes the characters after training the characters in vision assistance. but im not able to view the result string for more than 5 seconds. i really want to have some support regarding this issue. i've attached the program what i've used with this message. plz do help me with this issue asap.
Attachments:
OCR First Example with integrated2.vi ‏120 KB

Rudegar,
Our network is very stable and we aren't seeing any problems with anything else and besides, these worked fine until we did the update.
The wifi access point is in the same room about 6-8 ft. away and it's kind of hard to connect an appletv by ethernet to an iPad.
The appletv runs fine using it's built in applications, like Netflix, Disney, Radio, etc.
All of this makes me think it's not a wireless problem, but it's something in the appletv software after doing the update that won't hold the iPad connection.
I appreciate your suggestions and will pass your information along to our network person.
Any other ideas?
Thanks

Poor printing results

Pesky problem: prints too dark and colors lack saturation.
Have been printing on an Epson Artisan 50 with good results. Have added a new R2880. Both print with the same poor results NOW.
Use OS Mavericks, a 2-year-old iMac calibrated with Color Munki, custom ICC profiles for both printers, and PS 2014. Workroom is fairly dark with medium tan walls and 2 ‘Ott’ lights behind monitor.
Have made the following adjustments to see if I can find a solution:
•   With monitor have tried luminance of 80 and 120.
•   With monitor have tried white points of 55 and 65.
•   Have let PS and printer handle color.
•   Have converted color space from ProPhoto to sRGB and back again.
•   Have used 2 different packs of Epson Luster- both new.
I do view prints under the daylight corrected Ott lights and I take them outside in the sunlight to see if they're dark. And they still are! Wondering if somewhere I’m making a simple rookie mistake? I’ve followed both J. Schewe’s and the Digital Dog’s instructions as best I know how. How do you check that the driver settings are correctly configured? How do you check the output profile setting?
Any help is much appreciated.
Richard

Aperture really wants to manage colour during printing.
One common problem seems to be having your printer (driver) set to colour management, and then Aperture doing it as well. (This is probably not the only problem...)
One of the O'Reilly Inside Aperture Podcasts, with Joe Schorr, covers this very issue.

Poor Elgato results w/ 30" HD Display??

I have an Elgato 200 unit hooked to my G5 (dual 2.7, so plenty of power) with 30" HD display.
Though the Elgato unit works "ok", the picture quality is quite poor. I've thought that perhaps it's the same issue as when a non-HD signal is coming into my Plasma in the family room, but am seeking feedback to identify if I may have a setting that I can adjust to improve this.
I had been told by Elgato that I should get "at least" DVD quality, but it is not close to that. If I put a DVD into my PowerMac and play it, I get DVD quality and it looks good, so I know the 30" display can handle a "DVD-like" signal and produce a good picture.
Does anyone have experience and input that may help me here? Thanks!

I have a similar setup EyeTV 200 w/23" Apple Cinema Display. I'm going to assume your talking about viewing at full screen and that your feeding it a cable signal.
The quality your getting is normal. I never expected to view a cable source on this display with excellent results. I watch EyeTV at the Normal Size. You can't get DVD quality unless you put DVD quality in. Your comparison to your Plasma is exactly what is happening here.
I also have an EyeHome connected to a 32" Sony XBR HDTV. Watching recordings from the EyeTV 200 streamed to the EyeHome produces a very good picture. I would assume that watching the EyeHome on a standard definition TV would produce even better results.
Hope this helped out. Any other questions don't hesitate to ask.

Poor Benchmark results - WHY PLEASE??

Hi,
Can anyone help me with why I get a poor set of 3D Mark 2001se (330) results:7250 ???
The first purchase was a Nvidia GF4 Ti4200 XFX 128mb DDR. I initially tried this in my ECS K7S5A machine, and the benchmark result was just off 9000. Not Bad!
Then I purchase a MSI K7N2G-ILSR. I also upgrade the CPU from 1.4 T/Bird to 2.4XP
I spend weeks trying to get the VGA Drivers to install along with the nForce2 system drivers.
I've now updated the bios to V1.2, this I think was the main problem with the GF4. Although, when installing VGA and System drivers, I have to select each item from the device manager and individually update drivers. I wonder if the MOBO VGA need to be disabled somehow???
System Specs:
MSI K7N2G-ILSR
AMD Athlon XP PR2400+ Thoroughbred
Nvidia GF4 Ti4200 XFX 128MB DDR
512MB 333DDR (Nanya)
80GB WD Caviar (partitioned 20/10/50)
550W Q-tec PSU (I know - cheap !!!)
Volcano 9 HSF (Antec Silver Thermal Compound)
Win XP-PRO (SP.1)
Any help would be much appreciated!!
Markone.

Markone,
Update the motherboard and Vid card to the latest nVidia Drivers from the nVidia website.
Set the CPU to RAM Ratio to 1:1.
Tweak the drivers and see what scores you get.
Good Luck,
Richard

Lenovo G510 Camera Dark Poor camera Result Issue

Hi
i got new Lenovo G510 last week windows 8 64bit operating system core i3
In its specs it se clearly shown that it have 720 p hd camera but when i opened the camera app from start menu it only shows 0.1MP (4:3) and 0.3 (4:3) option and result is so dark and laggy
I checked the driver in device manger it shows properly installed by this name 'Lenovo Easy Camera" driver details are Microsoft driver version 6.2.9200.16664
i tried to update it from internet but it says its updated
i tried manually install the driver from here " C:\drivers\Camera Driver (AVC, Liteon, Bison, Chicony)" but Warning message pops up " NO APPROPRIATE DRIVER TO B INSTALLED"
Then i tried to install it from her C:\drivers\Camera Driver (AVC, Liteon, Bison, Chicony)\HD\Bison\3.13.301.1
but setup gets "failed" at the end and says "please plugin the device and restart the setup"
PLEASE HELP ME PLEASE

hai13h wrote:
I bought g510 week ago and the camera is terrible black, and blur i tried many drivers many ways but nothing succed... how to fix it!

OCR resulting in Grainy Images and junk OCRd Data

Why is OCRing not working on my PDF? I have two PDFs produced by two different Tiff to PDF tools. Both produce seemingly good PDFs that Acrobat reader has no problem rendering. The process we use to OCR the PDF is to load it into Adobe Acrobat X and to run an action that has a single step - Recognize Text (using OCR). The Primary OCR Language is English (US) and the PDF Output Style is Searchable Image. The PDF from the old tool can be OCRd in Adobe Acrobat X without any problems (the image looks good and the OCRd text is suitable). It appears that Adobe deskews the images automatically. The PDF from the new tool cannot be OCRd in Adobe Acrobat X without incurring some problems (the image looks grainy, and the OCRd text is junk). It appears that Adobe DOES NOT deskew the images automatically. To overcome the grainyness of using this second tool, we've attempted to turn on the PDF Output Style of "Searchable Image (Exact)". This overcomes the grainyness issue, but doesn't help with the junk OCRd text at all.

Here's a copy of some sample PDFs (one from each of the two processes/tools)
https://www.evernote.com/shard/s88/sh/d9103207-13a4-41fd-8b8b-694630899556/84fc679ef050329 2c272310d4248c16e
The file size for the old tool is smaller than that of the new tool.
Message was edited by: tomcruise1981

Poor PDF results from FreeHand MX v.11

When I export to PDF using Shift-Command-R, the resulting
effects (glows, drop shadows, etc.) don't render as they do within
FreeHand - they appear in the PDF as un-graduated transluscent
color blocks (everything else in the export, including placed
images and type, look great). When I try an alternative - printing
to Adobe PDF - the effects look okay, but wherever they overlap
something else in the drawing it creates a shift in the edge of the
overlap, fills offset from the edges of their shapes, and type
that's been converted to paths looks terrible...
Anyone have ideas about how to get a better result?

> When I export to PDF using Shift-Command-R, the
resulting effects (glows, drop
> shadows, etc.) don't render as they do within FreeHand -
they appear in the PDF
> as un-graduated transluscent color blocks (everything
else in the export,
> including placed images and type, look great). When I
try an alternative -
> printing to Adobe PDF - the effects look okay, but
wherever they overlap
> something else in the drawing it creates a shift in the
edge of the overlap,
> fills offset from the edges of their shapes, and type
that's been converted to
> paths looks terrible...
Raster effecs work only in low resoltion rgb output and they
do not work with cmyk colors. Raster effect resolution can be
changed but there are no other options. Raster effects have
connection tp color management settings while outline objects do
not. This of course makes about impossible to make the colors
match.
Also PDF export is unreliable and should not be used in print
production.
> Anyone have ideas about how to get a better result?
Create the effects with another program like Photoshop or
Illustrator.
Jukka
PS Hopefully Future version of FreeHand have the same effect
engine as Illustrator. Pdf export in Illustrator, however, do not
work properly either and Distiller must be used as well.

Poor 3DMark03 Results...help SOS

Well, I just built this computer and I got some crappy results I think
AMD 3000+ 64 90mm
MSI Neo2 Platinum
(2)512 PNY 3200 RAM running in dual channel
PNY GeForce 6800 unlocked
I scored 7495 on 3DMark03
I think it should be higher.
Does anyone know what is wrong? :-\
I am using bios 1.3 and have the latest Nvidia drivers :(

should get 11000 11500 with it set up ok
try w/ fast writes off if you get crashes but on will sometimes give 50 / 60 more
also same with side band addressing ive found
driver selection makes a hell of a difference
go here to get drivers http://www.guru3d.com or here http://www.tcmagazine.info/modules.php?op=modload&name=Downloads&file=index
they got most usually a day or to after they are ripped or appear mostly beta of course
if you want futuremark approved whql drivers then 66.93 are the latest most peole get good scores with 66.70 or 66.72 tho.
gcard bios makes a difference you can flash to diferent makes quite easily ive tried 5 different to get what i like now also gcard mhz is more important than ever nowadays go for 3dmark 05 i would tho 03 is still damn good and more an al over system bmark than 05
ram timings w/ an amd proc make a fair dif with 3dmark try 3.3.3.10 !!!! for benchmarking see what you get.
also disable any internet connection first and any firewall and virus protection close all unnecesary applications and defrag hd first remember to enable your security features before posting your scores tho!!!
you should get a fair improvment if you follow this !!!!
i gain 1000 points in 3dmark 05 by trying these things but it is trial and error !!

Poor quality results

I have Sony DCR-SR82 (HDD) camcorder, which uses MPEG-2 file format. Imported this directly into iMovie08 (cool!) and used iDVD08 to burn DVDs. It looks blurry. At least, not as sharp as video played directly from the camcorder on TV. So, I decided to explore some other avenues, ... namely using iMovie HD.
Here is the workflow:
MPEG-2 --> DV (MPEG Streamclip) --> iMovie HD6 --> iDVD08 (export from iMovie HD directly) --> burn DVD
This resulted in significantly sharper picture then those using iMovie08, but I am still not pleased with the quality. The edges are chopped up (like stairs), whenever there is a bit of a motion. This is really annoying. In iMovie HD6, while editing, it looks better, … Perhaps I am doing something wrong in the workflow described above.
Other info:
*Camcorder: Recorded in 16:9
*MPEG Streamclip conversion to DV : Kept all default settings (don’t really understand what they are)
*iDVD: Professional Quality.
*Media: DVD-R, Sony
Any help is highly appreciated!

Milan 011 wrote:
...How on Earth is this possible?
iMHD6 is a DV-editor for TV delivery.. so, the used codec CREATES interlacing to fullfill the international standards for NTSC/PAL ..
e.g.: you import stills (=no interlace), add a disolve= interlacing..
another example: you import 'progressive' material (special camera, CGI, whatever..) = any iM dv-stream export is now interlaced..
the capabilities of your playback device have no connections to the input.. any DVDplayer delivers interlaced.. or, you have to 'tell' it not doing so.. any 'tube' TV does interlacing..
commercial films on DVD are +per se+ progressive (=film), the DVD is usually in a video/TV standard = interlaced.. but both 'fields' of a progressive source show same content, even interlaced..
.. complicated issue to understand AND to explain in a foreign language..

QuickTime exports: poor quality result

This ties in with troubles I've been having with EyeTV exporting in bad quality. Since it uses the QuickTime exporter, I've narrowed my problem to QuickTime giving very bad quality exports. I know it is definitely QuickTime and not my settings, since I use the exact same settings in VisualHub and it converts the file beautifully.
This is become more of an issue because I was wanting to export has H.264 directly from Final Cut Pro, but since it too uses the QuickTime export it provides me with very poor clarity.
My system details:
Mac OS X 10.5.4, QuickTime 7.5.
I have Flip4Mac and Perian codecs installed for QuickTime. What is the issue here?

It's just a normal computer monitor. But the thing is if I a-b the exported video compared to the video before it was ever dropped into final cut pro, there is an obvious difference. So, I didn't think the monitor was an issue although I am sure it doesn't help. But, in the end the videos I do end up getting converted to flash and put on the website. Perhaps I should deinterlace?
G5 Mac OS X (10.4.9) Panasonic PV-GS300 3CCD

Poor Search Results

Hi,
I've been interested in a series of CD's with the basic title of: "Mexican-American Border Music". There are about 6-8 volumes.
When I enter the basic search all I get is Vol. 1. I have to jump over to Amazon and get exact titles and add these in to find additional offerings. iTunes offer most of the series yet they won't show up unless I search only the sub title or name of the volume, even though they all begin with the title: "Mexican-American Border Music"
This is real in efficient...I wonder how many other titles I'm missing?
I understand the search incorporates a tag and if it's not there.....But gosh a by label search or some work around to see all the offerings of a particular label would be great! I'm very interested in MANY Arhoolie label recordings but having to cross reference and use multiple searches to find them (and I find many after multiple tries of different terms!) is sending me to the record store to buy cd's instead.
Anyway or plug ins to "fix" this?
Any hints?
I'm relativley new to iTunes Store..
Thanks,
Whatrix1

Sorry, but there's no way for you to fix this on your end. The search functionality is entirely part of the iTunes Store, not of the iTunes application on your end, so no plugin or update could fix it.
The iTunes Store Power Search does seem to have problem picking out words in the middle of a name, particular when there's hyphenation involved. If you search, though, for "Conjunto" or "texas-mexican", then all of the albums should show up. You just have to try various variations of your search; initial words almost always work better, or entire phrases when the words are separated by a hyphen or other punctuation mark.
Regards.

BUG in OCR in Acrobat 9.3 on MRC (Mixed Raster Content) PDFs

I have lots of (non-searchable) PDFs, that were generated from scanned images (1 image per page @ 400 dpi) using LuraDocument from www.luratech.de.
The images are stored internally in the PDFs as MRC (Mixed Raster Content), that means, the PDF contains a foreground and a background layer for every image/page. These layers have a low resolution, are highly compressed and merged together by Acrobat (while displaying) using a high resolution mask layer. This results in very low file sizes, about 50 kB / page.
I'd like to make these PDFs searchable, but WITHOUT manipulating or changing the original image layers in any way. OCR software like FineReader or Omnipage seems to store images always with own algorithms, so that the image quality would suffer from the conversion and the size of the PDFs would rise significantly. Acrobat on the contrary offers to maintain the original image(layer)s by using the output style "Searchable Image (Exact)" in the OCR window. Now the problem:
After starting OCR, Acrobat applies OCR only at the first page (with good results) and deletes (!) all content (the image layers) on all other pages. For my eyes this seems to be a bug.
I tried a workaround: In Acrobat's Layers Panel I choose the menu option "Flatten Layers". Starting an OCR now, Acrobat does OCR on all pages of the PDF, but the OCR result is a disaster, less than 10% correct. Presumably Acrobat does not take the resulting (actually displayed) page content as input for its OCR, nor the high resolution mask layer, but instead one of the (low resolution, highly compressed) image layers described above.
Has anyone made similar experiences with MRC-compressed PDFs, e.g. PDFs generated by other MRC-Generators like JRAPublish ? Is there any workaround or bugfix ?
Thank You in advance !
L.Benic

I'm having the same issues. Using the latest version 9.3.3. Is this a bug? I tired calling adobe but their CR sounds like 3rd country only. Anyone can shed a light on this issue?

ClearScan - buggy OCR - arbitrary spaces in words

I've got 400 pages from a document, scanned @ 600 dpi color.With Acrobat 9 I converted them to 1 PDF, internally all pages stored as JPG2000, high quality. File size ~ 350 MB.
If I do OCR on this file (Acrobat 9 or 10) without ClearScan, everything is ok.
If I do OCR on this file (Acrobat 9 or 10) with ClearScan, the text looks perfect. But: If I copy and paste text (to the Windows Editor), most words contain spaces, I get "Abcd ef g hijklmnop" instead of "Abcdefghijklmnop". Of course, that also means, I can not find "Abcdefghijklmnop", if I try to search for it. Hence, over 90% of the OCR result is useless - because not searchable.
Now, for test purposes I exportet 1 page (from the original 350 MB non-OCR'd PDF) to a new PDF. I did OCR on this 1-page-file (with Acrobat 9) with ClearScan, and the above problem did not appear: No arbitrary spaces, text was ok and searchable.
Did nobody ever watch this behaviour, although it concerns both Acrobat 9 and 10 ?

I think I found the solution to your problem. Just save the problematic OCR Clearscan document as Text (Plain). By just doing so, Acrobat Pro corrects and eliminates all the spaces between characters. Then you only need to save again the PDF document, and you will see (even if hard to believe) that it is already fixed. At least it worked very well to me!

Acrobat XI ocr: access hidden layer?

Hi all,
I have historic documents (German and English) that I want to OCR so that the text is searchable *without* changing its appearance. I've tried with previous versions of Acrobat where it did not quite work and thought I give Acrobat XI (Windows 7) a try.
I use "searchable image" in the correct language ("Clearscan" and "exact" are not useful here). The ocr'ed text is in a hidden layer. Since an old-fashioned font is used, the ocr result is expectedly faulty. So I need to correct those results which is where the problems arise. The little sub-menu allows me to look for "problem areas" which are then marked in red. The individual entries can then be corrected one by one. However, these changes do not always seem to be transferred to the hidden layer. This is evident either from trying a search for the term (ctrl f), or exporting to Word. Both yield the original, not the corrected, ocr result.
Second problem: Once I mark a problem area as solved, there is no way to access that word, other than starting all over again.
Third problem: The keyboard shortcuts in the submenu don't always work.
Improving the scan quality is no solution because some older characters have no equivalent anyway.
The only solution seems to me to access the hidden text in some way and edit directly there. I did not find any mentioning of that in Acrobat's help or the forums, however. So I expect it's still not possible?
(As an aside, I'd like to submit the problem to Adobe but don't know how)

What you're asking for is not currently possible with the native tools in the Acrobat Family, but could be done with plugins.
When you run Searchable Image on a file, the text objects created on the page have a tag that sets their text rending mode to 3 (which tells compliant PDF applications that they are invisible but selectable - in effect they have no active stroke or fill). Acrobat doesn't let you change the text rendering mode, but there are third-party 'COS editor' plugins (Google for vendors) which can change any tags in the file stream. As you don't have any 'real' text in the file, the pseudo-workflow would be to search and replace the "/Tr/3" tag with "/Tr/0" so the text is visible, do your editing stuff, then reset it.

Poor OCR results MG6250

Similar Messages

Maybe you are looking for