CN201191870Y

CN201191870Y - Mobile phone having OCR recognition function

Info

Publication number: CN201191870Y
Application number: CNU2008200207577U
Authority: CN
Inventors: 王爱磊
Original assignee: Individual
Current assignee: Individual
Priority date: 2008-04-25
Filing date: 2008-04-25
Publication date: 2009-02-04
Anticipated expiration: 2018-04-25

Abstract

The utility model discloses a handset, in particular to a handset with OCR identification function, which comprises a handset host, a cam and a main board hard disk. The handset is characterized in that an OCR chip which is integrated with an OCR identification system is arranged on the main board hard disk, the OCR identification system enters into the video-image fore-process, the character feature extraction and the comparison identification through the video-image input of the cam and corrects the false character by the artificial correction and outputs the result. The cam of the handset can store a plurality of 'shot' data in the handset, which provides convenience to read for people timely, further memorizing repeatedly and reaching the goal of getting hold of the knowledge. The utility model plants the OCR function into the handset, which is beneficial to read book and newspaper for people or to store the data of the plane media for students greatly.

Description

Mobile phone with OCR recognition function

(1) technical field

The utility model relates to a kind of mobile phone, particularly a kind of mobile phone with OCR recognition function.

(2) background technology

In daily life, we see that from newspaper, magazine many data have value for preservation very much, and common people do not have copier at hand, record and the especially trouble that seems with notes.At any time to consult in the mobile phone to be very easily if can be stored into it, but spell them in the mobile phone not only consuming time but also consume power with phonetic or handwriting functions!

(3) summary of the invention

The utility model provides a kind of mobile phone with OCR recognition function easy to use in order to remedy the deficiencies in the prior art.

The utility model is achieved by the following technical solution:

A kind of mobile phone with OCR recognition function comprises mobile phone main body, camera, mainboard hard disk, and its special character is: an OCR chip that is integrated with the OCR recognition system is installed on the described mainboard hard disk.

Mobile phone with OCR recognition function of the present utility model, described OCR recognition system enter image pre-treatment, character features extraction, comparison identification successively through the image input of camera, through the word correction that manual synchronizing will be admitted one's mistake, the result are exported.

OCR (Optical Character Recognition) optical character identification.The OCR Chinese meaning is discerned literal by optical technology exactly.This technology can make equipment come identification character by the mechanism of optics.The mankind discern many things with eyes, and its mode is exactly a kind of optics mechanism.

The utility model is added to the OCR function in the mobile phone, makes things convenient for the scholar to store data.The utility model utilizes the camera of mobile phone that some data " bat " are stored in mobile phone, said " bat " be utilize existing OCR system data logging in mobile phone.The OCR system notes data with the form of picture, but utilize its Automatic Editing Function that has that the form of data with text is stored in the mobile phone, and can browse at any time, be convenient for people to browse at any time, and then repetitive memory, reach the purpose of grasping this knowledge! After the utility model is successfully implanted mobile phone with the OCR function, people who reads newspaper helping greatly often reading and the data on numerous students' memory plane medium.

(4) description of drawings

Below in conjunction with accompanying drawing the utility model is further described.

Fig. 1 is a structural representation of the present utility model;

Fig. 2 is the structured flowchart of the utility model OCR recognition system.

Among the figure, 1 mobile phone main body, 2 cameras, 3 mainboard hard disks, 4OCR chip.

(5) embodiment

Accompanying drawing is a kind of specific embodiment of the present utility model.This embodiment comprises mobile phone main body 1, camera 2, mainboard hard disk 3, and an OCR chip 4 that is integrated with the OCR recognition system is installed on the mainboard hard disk 3; The OCR recognition system enters image pre-treatment, character features extraction, comparison identification successively through the image input of camera, through the word correction that manual synchronizing will be admitted one's mistake, the result is exported.

OCR (Optical Character Recognition) optical character identification.It belongs to a knowledge of pattern identification, and its purpose will allow computer know what it has seen on earth exactly, especially written historical materials.Because OCR is a technology with the discrimination tug-of-war, therefore how debug or utilize supplementary raising recognition correct rate is the most important problem of OCR.And the media difference that exists according to written historical materials, and obtain the mode difference of these data, just derive of all kinds, various application.

An OCR recognition system, its purpose is very simple, just to do a conversion to image, make the figure in the image continue to preserve, have form then interior data of form and the interior literal of image, become the mobile phone literal without exception, the literal that the storage capacity that enables to reach image data reduces, identifies can re-use and analyze, and also can save manpower and time because of the keyboard input certainly.Its handling process is as follows:

From the image to result, export, must extract through image input, image pre-treatment, character features, comparison is discerned, after the word correction that manual synchronizing will be admitted one's mistake the result is exported.Make introductions all round at this:

The image input: the data that needs OCR to handle must be transferred to image in the mobile phone by camera.Now the medium-to-high grade mobile phone produced of each big communication apparatus company generally all is equipped with camera, has and takes a picture and camera function.Along with the progress of science and technology, the resolution of camera will be more and more higher, thereby make imaging quality also more and more clear simultaneously, and this can improve reading of OCR system and automatic editing speed greatly.

The image pre-treatment: the image pre-treatment is in the OCR system, the maximum module of must dealing with problems, from obtain one be not black be exactly white binaryzation image, or GTG, colored image, process to independently going out literal image one by one all belongs to the image pre-treatment.Comprise image normalization, removed the image processing of noise, image rectification etc., and the file pre-treatment that separates with word of picture and text analysis, literal line.Therefore aspect image processing, all reached the ripe stage in principle and technical elements, on the market or many available chained libraries are arranged on the website; Aspect the file pre-treatment, then with each tame ability; Image must be separated picture, form and character area earlier, even the layout direction of article, the outline and the content body of article can be distinguished, and the font of the size of literal and literal also can judging as original document.

Character features extracts: single with discrimination, and can the say so core of OCR of feature extraction, with what feature, how to extract, the direct quality discerned of influence, so also study the initial stage at OCR, the research report of feature extraction is many especially.And the chip that feature can be said so and be discerned, easy differentiation can be divided into two classes: a feature for statistics, as the ratio of counting of the black/white in the character area, when literal field is divided into several zones, this regional one by one black/white count than associating, just become a numerical value vector in space, when comparison, basic mathematical theory just is enough to deal with.And the another kind of feature that is characterized as structure, after literal image graph thinning, obtain the stroke end points of word, the quantity and the position in crosspoint, or be feature with the stroke section, cooperate special comparison method, compare the many methods of the recognition methods of hand-written Input Software on the line on the market based on this kind structure.

Comparison database: after input characters has been calculated feature, no matter be feature with statistics or structure, all must there be a comparison database or property data base to compare, the content of database should comprise the word collection literal of all desire identifications, according to the feature group of the feature extraction method gained the same with input characters.

Contrast identification: this is a module can giving full play to the mathematical operation theory, according to different features, select different mathematical distance functions for use, more famous comparison method has, the comparison method of theorem in Euclid space, lax comparison method (Relaxation), dynamic routine comparison method (Dynamic Programming, DP), and the database of neural network is set up and comparison, HMM (Hidden Markov Model) ... etc. famous method, for the result that makes identification more stable, also there is so-called expert system (Experts System) to be suggested, utilize the different complementarity of various feature comparison methods, make the result who identifies, its confidence degree is high especially.

The words reprocessing: because the discrimination of OCR and can't reaching absolutely, or want correctness and the confidence value strengthening comparing, some debugs or even the function of help corrigendum, also become necessary in an OCR system module.The words reprocessing is exactly an example, utilizes in identification literal and its possible similar candidate's sub-block after the comparison, finds out the most logical speech according to the identification literal of front and back, does the function of corrigendum.

Word database: the dictionary of being set up for the words reprocessing.

Manual synchronizing: a good OCR software, except a stable image processing and identification core are arranged, reducing outside the error rate, the operating process of manual synchronizing and function thereof also influence the treatment effeciency of OCR.

Result's output: output is the simple thing of part in fact, but must see the user with OCR on earth for what? someone is as long as text is made the usefulness that re-uses of segment word, so as long as general text file, someone will be beautiful bright with the input file striking resemblances, so the function, the someone that have original text to reappear pay attention to the literal in the form, thus will with software combination such as Excel.No matter how to change, all just export the variation of File Format.

Claims

1. the mobile phone with OCR recognition function comprises mobile phone main body (1), camera (2), mainboard hard disk (3), it is characterized in that: an OCR chip (4) that is integrated with the OCR recognition system is installed on the described mainboard hard disk (3).

2. the mobile phone with OCR recognition function according to claim 1, it is characterized in that: described OCR recognition system is through the image input of camera, enter image pre-treatment, character features extraction, comparison identification successively,, the result is exported through the word correction that manual synchronizing will be admitted one's mistake.