|
|
|
|
|
Mobile Reader is a software device to
extract text information through mobile
image processor and OCR engine after analyzing
the image taken by the digital camera on
celluar phone or PDA.
|
|
|
 The
first commercialized Asian OCR
engine for the camera phones
in the worlds
|
|
|
Mobile Reader is the first
commercialize software which
can extract text information
in limited hardware environment
and now deliver to camera phone
manufacturers, such as Samsung
Electronics, LG Electronics
and Pantech&Curitel for
business card reader and OCR
for e-Dic.
|
|
|
|
High
performance in limited memory
and digitial camera for mobile
device
|
|
|
To optimize on mobile devices,
Mobile Reader minimize memory
usage, about 650KB to 1.9MB(650KB:
Alpahnumeric and symbols, 1MB:
when including Korean, 1.9MB:
when including Korean or Chinese)
and it offers high recognition
rate on digital camera.
|
|
|
|
Major
Functions are
|
- Digital camera image
enhancement
- Structure analysis and
extracting text area
- Character recognition
- Recognition post-processing
- Offering useful user
interface module - select
word, score image focus
in real-time, extract partial
image and etc.
|
|
|
|
|
|
|
|
|
|
|
|
Major Characteristics
|
|
- Unlike scanning environment, camera
is very sensitive to the environment,
and there may happen many problems according
to distortion, light intensity, distance,
focal point, etc. Mobile Reader, developed
based on its accumulated technology
for over 10 years, has overcome these
problems and makes it possible to extract
information from the image taken by
a camera.
- With limited memory of mobile device,
it functions in the low-definition camera
environment of processing
- Supporting multi-language recognizer
for Korean, English, Chinese, and European
languages.
- As it was developed with standard
development language to be applied to
diverse OS and H/W, it shows high transplant
rate.
-
Supporting self memory management function
- Removing
floating point operation for supporting
mobile processor, Using only integer
operation for all processes
- Secured
4-byte alignment in all memory reference
|
|
|
Functional Characteristics
|
|
- About irregular illumination and
image quality reduction that may happen
from camera image, Mobile Reader
improves the image quality and advance
inarization by means of image processing.
- Extremely small size OCR engine
which is optimized for the limited mobile
environment.
- Offering useful user interface module-select
world, score image focus in real-time,
extract partial image and etc
|
|
|
Mobile Reader won IR52 Jangyoungshil
Award
|
|
|
|
|
‘Mobile Reader’ awarded
on 49th week of year 2005
IR52 Jangyoungshil Award
is the best-recognized award
in industrial technology of
Korea, co-hosted by KITA and
Maeil Newspaper, and supported
by MST.
(
http://www.ir52.com/)
|
|
|
|
|
|
|
|
|
|
|
|
Main Functions
|
|
|
|
|
Function
|
Description
|
|
Image Process
|
Image enhancement
|
|
Image improvement
|
|
Binarization
for high performance
whit extremely high
speed
|
|
Document Analysis
|
Document structure
analysis
|
|
Text field extraction
|
|
word field extraction
|
|
Character Recognition
|
Korean, Chinese
, Englis
|
|
European language
(Spanish, French,
German, Italian.)
|
|
Post-process
|
Name field extraction.
|
|
Auto compensation
for simple recognition
error
|
|
Chinese character
--> Korean :
auto conversion
|
|
Image Format
Conversion
|
Non-compression:
RGB, YUV, BGGR and
Gray
|
|
Compression :
JPEG (JPEG decoder
optional)
|
|
UI Support
|
Partial image
extraction, Zoom-in/out
|
|
Document structure
analysis for supporting
character rows and
word selection UI.
|
|
Score image focus
in real-time to
improve capturing
convenience for
user
|
|
|
|
|
|
|
H/W Spec
|
|
|
|
|
Camera
|
1M pixel or higher
|
|
5~8.5cm macro
mode (recommending
Auto Focus)
|
|
Processor
|
ARM9 or comparable
speed
|
|
Memory
|
ROM : approx
650KB ~ 1.9MB
|
|
RAM : approx
500KB (at semi-auto
recognition, exclude
of input image)
approx
1.5MB (at auto recognition,
exclude of input
image)
|
|
OS
|
OS independent
(irrelevant to OS
- Engine product)
|
|
|
|
|
|
|
|
|
Business card recognition, OCR
for e-Dictionary
|
|
|
|
|
|
|
|
|
|
|
|
|
|
