CN1617163A - Method for selecting treating object in character identification of portable terminal and portable terminal - Google Patents
Method for selecting treating object in character identification of portable terminal and portable terminal Download PDFInfo
- Publication number
- CN1617163A CN1617163A CNA2004100889727A CN200410088972A CN1617163A CN 1617163 A CN1617163 A CN 1617163A CN A2004100889727 A CNA2004100889727 A CN A2004100889727A CN 200410088972 A CN200410088972 A CN 200410088972A CN 1617163 A CN1617163 A CN 1617163A
- Authority
- CN
- China
- Prior art keywords
- character
- image
- portable terminal
- recognition
- terminal device
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M1/00—Substation equipment, e.g. for use by subscribers
- H04M1/02—Constructional features of telephone sets
- H04M1/0202—Portable telephone sets, e.g. cordless phones, mobile phones or bar type handsets
- H04M1/026—Details of the structure or mounting of specific components
- H04M1/0264—Details of the structure or mounting of specific components for a camera module assembly
Abstract
The present invention relates to a processing object selective method in the character recognition processing of the portable terminal and the portable terminal thereof. A portable information terminal with a camera wherein accurate character recognition processing requires that a character string to be recognized not be inclined in an image, and if there are two character writing directions, that is, vertical writing and horizontal writing, correct character recognition requires that a character line direction be specified accordingly, which both impose a substantial burden of specification or correction on a user, and a search using as the key a recognition result of character recognition of Japanese, which unlike English has no character breaks, imposes a substantial burden of search word specification on a user. For an appropriate correction of an inclination of a character line, an indicator showing a character line inclination is displayed on a screen of an information terminal device. For a search using as the key a recognition result of character recognition of Japanese, a search word is specified by means of morphological analysis results and cursor position information.
Description
Technical field
Process object system of selection when the present invention relates in portable information terminal, carry out the optical profile type character recognition.
Background technology
At the portable information terminal that image input functions such as camera have been installed, the technology of the image of taking being implemented the optical profile type character recognition is developed.But with the captured image of camera of portable information terminal, owing to reasons such as hand swings, with respect to picture, the situation that character string is taken at a slant is more.Therefore, when character identification result mistake (misreading), or the angle that allows the user adjust when taking takes once more, or uses input media correction recognition result such as keyboard.
As the technology of revising the inclination of identifying object character string before handling in identification, in patent documentation 1, disclose detect charged to the identifying object character string with paper on the good mark more than 2 of record in advance, according to resulting inclination thus, the technology of discerning after being rotated is automatically attempted.
In addition, in non-patent literature 1, disclose at portable information terminal identification English word and searched the such using method that combines character recognition and dictionary retrieval of Britain and Japan's dictionary.
Patent documentation 1: the spy opens flat 11-250179 communique (4～7, the 3rd figure)
Non-patent literature 1:H.Fujisawa, H.Sako, Y.Okada, and S-W.Lee, " InformationCapturing Camera and Developmental Issues; " In Proc.Int.Conf.DocumentAnalysis and Recognition, ICDAR, 99, Bangalore, India, Sep.20-22,1999 ' pp.205-208.
Handle in order to implement character recognition accurately, importantly the character string of identifying object does not tilt in image.But, when the information terminal device shooting digital pictures of the portable terminal device that uses digital camera or band camera etc., the situation that fixes this information terminal device with hand is more, for the character string that makes identifying object does not tilt in image, needs special note (first problem) when taking.
In the function of the optical profile type character recognition of carrying out for the captured image of portable terminal device that uses the band camera-enabled, when character string is being tilted shooting, exists in the prior art and can not carry out character recognition, or do not reach the problem of enough accuracy of identification.Therefore, existed for and obtained correct character identification result, must take the problem of the character of identifying object once more.
Though in patent documentation 1, record the technology that detects inclination by the mark of putting down in writing on paper more than 2, this technology must identifying object with paper on stamp or charge to mark in advance, impracticable when identification business card etc.In addition, when the image taken being implemented character recognition handle, above-mentioned such special attention takes so long as not paying, the situation that just exists the identifying object character string to tilt in image.At this moment, implementing to use before character recognition is handled image processing software etc. to carry out the angle modification (second problem) of image.
In addition, when the record direction of character is write across the page and erected when writing 2 kinds of literary styles,, when character recognition, also must set the identifying object character string and be to erect and write or write across the page even after the angle of adjusting image, extracted the identifying object character string.When the image of obtaining business card etc. and when implementing character recognition and handling,, there is the each problem that all must set recognition mode owing to there is the perpendicular form of writing and writing across the page.In addition, newspaper, magazine etc. perpendicular write and document that the character string of writing across the page is mixed in, switch the burden also big (the 3rd problem) of the record direction of character.
When the device of the portable information terminal that utilizes the band camera etc., can infer, the road of taking the vehicles through regular meeting is first-class, the utilization under the very difficult environment of taking with the position of image stabilization.But, in the prior art,, just can not get enough character recognition precision for image if the identifying object character string is not selected under the state that tilts to be suppressed in very among a small circle.Therefore, the user must pay special attention (the 4th problem) in order to adjust angle when taking the identifying object image.
Moreover, at the portable information terminal of band camera, when the result that will use character recognition carries out dictionary retrieval or network retrieval, when being English word, because be separated into word units, so select the word of searching object to be easier to by the space.But, if same processing is applicable to Japanese, because it is different with the situation of English, there is not the separation that causes by the such arrangement information in the space between word, so the user must select character ground of character of the character string of searching object, specify burden big (the 5th problem).
Summary of the invention
The objective of the invention is: in view of these problems, photographer's burden when alleviating the portable information terminal that uses the band camera and taking the character recognition object is provided, perhaps alleviates system or method the burden of the image correction after taking for the time to the preferable angle of character recognition.
In addition, the present invention also aims to, provide when the character recognition Japanese and when carrying out dictionary or network retrieval, can alleviate the user's of the character string of specifying searching object the system or the method for burden.
In order to solve above-mentioned first problem, show on the picture of information terminal device the inclined degree in the image of identifying object character string in real time oblatio give photographer's angle display.The user by taking in position, can shoot the image that is suitable for the character recognition processing while information mobile information terminal device or the identifying object thing of seeing that the angle indicator is shown.
In order to solve above-mentioned second problem, be provided at when the image of having taken is rotated, implement to become the function that the row of the character string of identifying object extracts in real time.The user passes through simple key operation etc., the image that rotation has been taken on the picture of information terminal device, the identifying object candidate character strings that real-time confirmation is extracted by row.By the time point that is extracted in desirable identifying object character string, user's processing of stopping the rotation selects when several rows are extracted to want that the character string of discerning carries out identification and handle, and realizes the easy that the character recognition of the image taken is handled.
Automatically judge that in order to solve above-mentioned the 3rd problem, to provide this identifying object character string is perpendicular writing or the function of writing across the page.In the automatic judgement of this identifying object character string direction, use the asperratio of the boundary rectangle of the identifying object character string that is extracted.Specifically be that after the ratio of the height and width of the boundary rectangle of identifying object character string and setting compared, judgement was perpendicular write characters string or writes across the page character string and enforcement identification processing.When the direction-agile of the picture of the mobile communication terminal that uses, also can implement the perpendicular switching of writing/writing across the page automatically according to the direction of picture.
In order to solve above-mentioned the 4th problem, the angle of inclination of Tracking Recognition object character string is provided, automatically generate the device of the boundary rectangle that is used to select the identifying object character.Specifically be, the angle of inclination of using spy for example to open the method instrumentation identifying object character string of flat 7-141465 " slant detection method of document image ", making to be rotated into and make that the identifying object character string is a horizontal level with respect to image when writing across the page, is the image of upright position with respect to image during for perpendicular writing., generate the boundary rectangle of identifying object character string, make the rotation boundary rectangle image that is appended to this rotation back image thereafter.Then, rotate this rotation boundary rectangle image, turn back to the angle of inclination of original character string, show in the display device of information terminal device.
In order to solve above-mentioned the 5th problem, provide with lower device: promptly the result behind the character recognition Japanese is carried out the plain analysis of voice, generate the candidate character strings of searching object automatically, the user selects these candidates respectively or selects the combination of these candidate character strings.
By show the heeling condition of the image of identifying object in the mode of visually understanding easily, can allow user's clear understanding inclination take place to image, so just so that being handled suitable angle, character recognition comes photographic images easily.
In addition, about captured image of past, owing to can directly edit the image of inclination, and its result is implemented character recognition handle, therefore needn't take once more.
In addition, even under the angle modification situation of difficult, because can under the state that tilts, carry out the selection of character string, so, also can carry out character recognition and handle to having the image of inclination to a certain degree.
Description of drawings
Fig. 1 is the block diagram of the portable information terminal of the embodiment of the invention.
Fig. 2 is the process flow diagram of the embodiment of the invention.
Fig. 3 is the process flow diagram of the embodiment of the invention.
Fig. 4 is the key diagram of the angle modification of the embodiment of the invention.
Fig. 5 is the process flow diagram of the embodiment of the invention.
Fig. 6 is the diagram of the picture of the expression embodiment of the invention.
Fig. 7 is the process flow diagram of the embodiment of the invention.
Fig. 8 is the key diagram of the character string boundary rectangle generating mode of the embodiment of the invention.
Fig. 9 is the block diagram of the portable information terminal of the embodiment of the invention.
Figure 10 is the process flow diagram of the embodiment of the invention.
Figure 11 is the key diagram of the rectangle coordinate table of the embodiment of the invention.
Figure 12 is the selection mode key diagram of the retrieval candidate of the embodiment of the invention.
Figure 13 is the key diagram of the rectangle coordinate table of the embodiment of the invention.
Figure 14 is the key diagram of the selection region list of the embodiment of the invention.
Figure 15 is the key diagram of the rectangle coordinate of the embodiment of the invention.
Figure 16 is the diagram of the explanation embodiment of the invention.
Figure 17 is the key diagram of the rectangle coordinate table of the embodiment of the invention.
Figure 18 is the diagram of the explanation embodiment of the invention.
Figure 19 is the diagram of the explanation embodiment of the invention.
Figure 20 is the key diagram of the rectangle coordinate table of the embodiment of the invention.
Figure 21 is the diagram of the explanation embodiment of the invention.
Embodiment
Use Fig. 1～preferable a kind of embodiment of 20 explanation the present invention.Character recognition mode of the present invention goes for mobile information system that reads and discern by business card etc. etc., for example goes for the character recognition function of carrying out on mobile phone.
Fig. 1 is an example that is suitable for the block diagram of portable information terminal of the present invention.Have in this example: portable information terminal main body 100; As the business card of identifying object image-input device 110 with the camera of optical mode input or scanner etc.; The display device 120 of the image of demonstration identifying object or the CRT of character identification result, cursor 121 etc. or liquid crystal etc.; Disposed the input media 130 of button 131 grades that the user can operate; Be installed in the control part 140 in the terminal body for the control of carrying out portable information terminal integral body; And the character recognition portion 150 that carries out character row extraction 151, character recognition processing 152 etc.; Have the quantification function 161 of character row inclination and the image processing part 160 of image rotation processing function 162.
Fig. 2 is an example (first embodiment) of the character recognition of implementing to be suitable for the device that is used to solve first problem flow process when handling.The user is the OCR function at the beginning, and the animated image of importing from image-input device 110 just is displayed on the display device 120 (S201).Character recognition portion 150 carries out character row extraction processing (S202) to the cross zone of cursor 121 of waiting that has shown on the image display device 120 at once, shows the boundary rectangle (S203) that surrounds the character row that is extracted.
Simultaneously, by the inclination of image processing part 160 quantification character rows, and the value of this quantification carried out visual (S204) with the form of histogram etc. on angle display 123.The value of quantification is unrestricted, as long as reflected that the inclined degree of character row and image is just passable, if but for example adopt reciprocal proportional value of the angle θ that forms with character row and image end limit, then when degree of tilt is little, just on angle display 123, show big value, thereby the user can operate intuitively.
Processing turns back to step (S201), till the user presses (S205) shooting push button, below repeating (processing of S201～S204), and the continuous updating picture shows.
The user is with reference to mobile terminal apparatus or identifying object thing with angle display 123, presses shooting push button (S205) back in place and carries out image taking (S206).If then press recognition button (S207), then the character string in the shown boundary rectangle of step (S203) is carried out character recognition and handle 210, and show recognition result (S211).
After pressing shooting push button (S205), when delete button is pressed (S208), just deletes the image of having taken and turn back to step (S201).When in addition button is pressed, be transferred to operations necessary (S209) respectively.
Fig. 3 is an example (second embodiment) of the character recognition of implementing to be suitable for the device that is used to solve second problem flow process when handling.The user of portable information terminal carries out after the operation (S301) that captured rest image was packed in the past, and this rest image just is displayed on display device 120 (S302).Character recognition portion 150 carries out character row immediately and extracts (S303), boundary rectangle demonstration (S304) and angle display demonstration (S305).Handle the key input that is transferred to the user at this time point and wait for (S306).
When the user presses identification when carrying out button (S307), handle (S310) to carrying out character recognition immediately in the character string of the inside of the shown boundary rectangle of step (S304), and character display recognition result (S311).When the user presses arrow button (S308), according to the button of pressing, with image direction rotate to an angle (S309) to the left or to the right.At this moment, the center of rotation is the center of character row rectangle, but also can be by a bit being rotated processing as the center on the image of user's appointment.
When continuing to pin arrow button, image rotates continuously, and boundary rectangle shows also real-time update thereupon.Consider the easy to use of user, can function in an acting capacity of identification and carry out the operation that button is pressed with stopping operation (finger being removed) that arrow button pressing from button.
Fig. 4 is an example of the shown image of display part 120 in first embodiment and second embodiment.Be the animation imported from image-input device 110 in the first embodiment, second embodiment be before captured rest image, be presented at display part 120 as image 400.
Identifying object character string 401 in the image 400 tilts at this time point.Show tracking cross 402 in the central authorities of picture as rotation center.Character recognition portion 150 generates the boundary rectangle 403 that surrounds identifying object character string 401, and is presented on the picture.On angle display, show the histogram (404) of the inclined degree of the identifying object character string 401 that expression tilted.
Use hand-held portable information terminal (camera) (406) by rotating in the first embodiment, or rotate image shown on the picture by in second embodiment, operating arrow key (405).By the rotation of image, along with the inclination of identifying object character string 401 diminishes, the shape of boundary rectangle also changes (407) synchronously.
In addition, at angle display, the big value that the inclination of expression identifying object character string 401 is diminished shows (408) as histogram.The user is by carrying out the rotary manipulation of image repeatedly, and becomes big position in the value that angle display shows and carry out character recognition and handle, and can obtain the high character identification result of precision.
Fig. 5 is an example (the 3rd embodiment) of the character recognition of implementing to be suitable for the device that is used to solve first problem flow process when handling.Because (S501～S506) (S201～S209) identical is so omit explanation with step for step.
Be pressed in the recognition button time point of (S504) calculates the asperratio (ratios of height and width) of the boundary rectangle of identifying object character string, compares (S507) with the value α that predesignates.If asperratio is bigger than setting α, then be judged as perpendicular write characters string, implement (S510) such as parameter settings of perpendicular write characters string identification usefulness, and implement character recognition and handle (S511), display result (S512).
Equally, if asperratio and setting α are relatively big unlike setting α, then continue asperratio and setting β are compared (S508).If asperratio is littler than setting β, then be judged as the character string of writing across the page, implement write across the page (S509) such as parameter settings of character string identification usefulness, and implement character recognition and handle (S511), display result (S512).If asperratio, is then thought character string in the scope below the α and more than the β not by abundant angle correction, do not change identification over to and handle.
Fig. 6 is an example of the shown image of display part 120 in the 3rd embodiment.When for the character string 601 of writing across the page, boundary rectangle high 602 littler than wide 603.If the height that asperratio is defined as boundary rectangle is wide divided by boundary rectangle, then asperratio than 1 hour boundary rectangle for growing crosswise.
For example, when the setting β of handle and asperratio comparison is set at 0.5, start character recognition, then implement automatically as writing across the page the necessary setting of character string if be lower than in asperratio under 0.5 the state.Equally, when for perpendicular write characters string 604, boundary rectangle high 605 wide 606 big.
Under the definition of asperratio same as described above, then when asperratio was bigger than 1, boundary rectangle was for perpendicular long.For example, when the setting α of handle and asperratio comparison is set at 1.5, start character recognition, then implement automatically as the perpendicular necessary setting of write characters string if be higher than in asperratio under 1.5 the state.
Fig. 7 is an example (the 4th embodiment) of the flow process when implementing to be used to solve the character recognition processing of device of the 4th problem.
After the character recognition object images was transfused to (S701) from image-input device 110, image processing part 160 just calculated the angle (S702) to the image of identifying object character string immediately, and this angle part is revised in the character recognition object images rotation of being imported.About revising direction, if the character string of writing across the page then rotates to be horizontal direction with respect to picture, if perpendicular write characters string then rotates to be vertical direction with respect to picture.
Then, this rotation back image is implemented character string extract, additional boundary rectangle (S704) on this rotation back image is saved in frame buffer with this image.The image of preserving at frame buffer can be an integral image, also can only be additional boundary rectangle inside.
Then, should rotate the back image and just in time reverse, make the image with original input image inclination same degree, be presented at the display part 120 of end device by the detected angle part of step (S702).If the user does not carry out any operation, then return step (S701), new input picture is repeated the step (processing of S701～S707).
If press recognition button, then read in the image (S710) that step (S705) is stored into frame buffer, (S711), character display recognition result (S712) are handled in this fulfillment character recognition.
Fig. 8 is an example of the state of the image handled in the 4th embodiment.Identifying object character string 802 on the identifying object image of being imported from image-input device 110 801 is the state that tilts with respect to picture.Image processing part 160 detects the edge angulation 803 of this identifying object character 802 and picture, with the image anglec of rotation 803 just in time, makes the identifying object character string become level with respect to picture thereby revise, and makes rotation correction image 804.
For the identifying object character string 806 on the rotation correction image 804, character recognition portion 150 implements character row and extracts, and additional boundary rectangle.Image processing part 160 should rotate just in time reverse angle 803 of correction image 804, generated the image 807 that turns back to former identifying object image 801 equal angular, was presented at display device 120.
First to the 4th top embodiment can be distinguished a realization certainly, also can realize with the form that optionally adopts all or part of.
Below, use Fig. 9 to 12 pair of the 5th embodiment that is used to solve the 5th problem to describe.Fig. 9 is an example of block diagram that has been suitable for the portable information terminal of the 5th embodiment.With the difference of Fig. 1 is to have appended retrieval language extracting part 170, e-dictionary 171.
Figure 10 is an example of the flow process when the device of Fig. 9 implements to be suitable for character recognition, the dictionary retrieval process of the device that is used to solve the 5th problem.The character recognition object images is after image-input device 110 inputs (S1001), after 160 pairs of original images of image processing part are implemented appropriate image processing, by the character row extracting part 151 extraction character rows (S1002) of character recognition portion 150.Afterwards, for the character row that is extracted, character is isolated by per 1 character in character row identification part 152, and the result (S1003) of output identification.In recognition result, include the character code of per 1 character and corresponding therewith rectangular coordinates.
The result of character recognition is transfused to retrieval language candidate extracting part 170, by the plain morpheme (S1004) that continuous character string is decomposed into word etc. of analyzing of voice.For example, when the display of the portable information terminal of Fig. 9 show be the such character string of " grammatical Zhi Knowledge The makes う と " (" if using knowledge of grammar ") time, generate table 1100 as shown in Figure 11.Store by the plain character string and the corresponding therewith rectangular coordinates of being decomposed of analyzing of voice.
Use the data of table 1100, on the display of portable information terminal, show the candidate (S1005) of searching object.For example, the relatively centre coordinate of cursor and the candidate rectangular coordinates of table 1100 will comprise that the rectangular coordinates of candidate of the centre coordinate of cursor shows on display with as shown in figure 12 form.
Then, by pressing cursor movement key 174 or 176, the rectangle of mobile search object word shows as 1201, when having shown the rectangle of wanting to retrieve by select button 175, decision searching object word (S1006).The searching object word that is determined is exported to e-dictionary portion 171.In e-dictionary portion 171, the searching object word imported as keyword retrieval e-dictionary (S1007), and is presented at (S1008) on the display with result for retrieval.
Though use the plain candidate that has generated searching object of analyzing of voice in the present embodiment, but, also can be following method: the place that promptly becomes the classification variation of characters such as " hiraganas " in the character string of recognition result from " Chinese character " disconnects, and generates the method for candidate.Perhaps can judge with the geometric information such as place of the variation of space or character boundary in conjunction with character class.
Below, use Fig. 9, Figure 10 and Figure 13 to 15, the 6th embodiment that is used to solve the 5th problem is described.In the present embodiment, relate to the situation that the Chinese character row that are made of a plurality of morphemes are arranged as " grammatical Zhi Knowledge The makes う " (" the use knowledge of grammar "), suppose that cursor is arranged in any 1 of " grammatical Zhi Knowledge " (" knowledge of grammar ") character string.
Owing to exist the user only to want as " syntax " (" grammer ") or “ Zhi Knowledge " morpheme (" knowledge ") is as the situation of searching object and think the situation of the compound word integral body that retrieval " grammatical Zhi Knowledge " (" knowledge of grammar ") is such, therefore the following describes the processing that alleviates these selection burdens.In the 6th embodiment, because be that candidate extracts (S1004) and candidate shows (S1005), candidate selection (S1006), so only this processing is described with part different in the treatment scheme of Figure 10.
In candidate extracts (S1004), by with identical processing shown in the 5th embodiment, generate candidate by plain analysis of voice, generate corresponding therewith rectangular coordinates table 1300 as shown in figure 13.Then, generate by table 1300 be used for selecting " syntax ", " grammatical Zhi Knowledge ", “ Zhi Knowledge respectively " the area coordinate table 1400 of (" grammer ", " knowledge of grammar ", " knowledge ").Enter this zone if this table is used for the cursor centre coordinate, then show the rectangle of the candidate corresponding with it.
Express to Figure 15 pattern the X coordinate of the rectangular coordinates of this table.With " syntax ", " grammatical Zhi Knowledge ", “ Zhi Knowledge " (" grammer ", " knowledge of grammar ", " knowledge ") corresponding respectively selection zone is 1500,1501,1502; according to involved which zone of the centre coordinate of cursor, from table 1400, select the rectangular coordinates that shows as the retrieval candidate.Then selected rectangular coordinates is presented at (S1005) on the display.
Figure 16 represents the example of shown rectangle.(a) be that cursor is positioned at “ Zhi Knowledge " demonstration example during the selection of (" knowledge ") zone; (b) being the demonstration example of cursor when being positioned at the selection zone of " grammatical Zhi Knowledge " (" knowledge of grammar "), (c) is the demonstration example of cursor when being positioned at the selection zone of " syntax " (" grammer ").When the user is shown when the candidate of wanting to retrieve, presses options button and select searching object word (S1006).Present embodiment is selected to be illustrated to the character string in 1 character row, but by holding in the lump the area coordinate table of character row to greatest extent, can relate to the selection of a plurality of character rows.
Below, use Fig. 9, Figure 10, Figure 13, Figure 17, Figure 18, the 7th embodiment that is used to solve the 5th problem is described.The same with the 6th embodiment, relate to the situation that the Chinese character row that are made of a plurality of morphemes are arranged as " grammatical Zhi Knowledge The makes う " (" the use knowledge of grammar "), suppose that cursor is positioned at the situation of " syntax " (" grammer ") part.In addition, the same with the 6th embodiment, only illustrate that candidate extracts (S1004) and candidate shows (S1005), candidate selection (S1006) part.
In candidate extracts (S1004), by with identical processing shown in the 5th embodiment, generate candidate by plain analysis of voice, generate corresponding therewith rectangular coordinates table 1300 as shown in figure 13.Then, generate respectively and morpheme and the corresponding rectangular coordinates table 1700 (Figure 17) of its compound word from table 1300.Suppose that table is with upper left point coordinate ordering.
Show in (S1005) that in candidate as shown in figure 18, the centre coordinate of initial display highlighting is comprised in the rectangle (1800) of the morpheme of its rectangular area.Then, press cursor key 176 at every turn, demonstrate the rectangle of the following table 1700 that is sorted.The user can press options button 175 at the time point that the rectangle of wanting to retrieve is shown, retrieve electronic dictionary (S1006).
Below, use Fig. 9, Figure 10, Figure 13, Figure 19 that the 8th embodiment that is used to solve the 5th problem is described.The same with the 6th embodiment, relate to the situation that the Chinese character row that are made of a plurality of morphemes are arranged as " grammatical Zhi Knowledge The makes う " (" the use knowledge of grammar "), suppose that cursor is positioned at the situation of " syntax " (" grammer ") part.In addition, the same with the 6th embodiment, only illustrate that candidate extracts (S1004) and candidate shows (S1005), candidate selection (S1006) part.
In candidate extracts (S1004), by with identical processing shown in the 5th embodiment, generate candidate by plain analysis of voice, generate as shown in figure 13 the rectangular coordinates table 1300 corresponding with it.Show the candidate rectangle (1005) that includes the cursor centre coordinate with as shown in figure 19 1900 form then.
At this, when wanting to select " grammatical Zhi Knowledge " (" knowledge of grammar ") such compound word, press " 1 " key of the meaning that is endowed the starting point of specifying range of choice after, press cursor movement key 176, show the rectangle as 1901.If press cursor movement key 176 again,, select the zone extended then as 1902.The user presses options button 175 using cursor key to demonstrate the time point of suitable searching object word, selects searching object word (S1006).
Below, use Fig. 9, Figure 10, Figure 20, Figure 21 that the 9th embodiment that is used to solve the 5th problem is described.Relate to the situation that the Chinese character row that are made of the morpheme more than 3 are arranged as " grammatical Zhi Knowledge handles The " (" knowledge of grammar processing "), suppose that cursor is positioned at “ Zhi Knowledge " (" knowledge ") situation partly.
When for the Chinese character row that are made of the morpheme more than 3, the user usually or consider the Chinese character row wholely as searching object or thinks only to retrieve the morpheme that includes the cursor centre coordinate, selects this frequency of two kinds higher.Therefore, the following describes the processing that alleviates these selection burdens.In addition, the same with the 6th embodiment, only illustrate that candidate extracts the part of (S1004) and candidate demonstration (S1005), candidate selection (S1006).
In candidate extracts (S1004), by with identical processing shown in the 5th embodiment, analyze when generating candidate by voice is plain, generate storage and candidate corresponding characters kind and rectangular coordinates, table 2000 as shown in figure 20.In this said character kind being, is exactly " Chinese character " if belong to " Chinese character ", is exactly that " hiragana " is such if belong to " hiragana ", means by character kind sorting result.Then, merge the rectangle have with the candidate of candidate (morpheme) the identical characters kind that includes the centre coordinate of cursor, and show (1005) with form as shown in Figure 21.
At this, the Chinese character row are whole and think only to select “ Zhi Knowledge when not wanting to select " during (" knowledge ") this candidate, press and be endowed " # " key that switches the meaning of preference pattern, and show rectangle as 2101.Moreover, when wanting to select “ Zhi Knowledge to handle " during (" knowledge processing ") character string, after pressing " 1 " key that is endowed the meaning of specifying the range of choice starting point, press cursor movement key 176, show 2102 such rectangles.The user uses such key operation, presses options button 175 at the time point of the rectangle that shows suitable searching object word and selects searching object word (S1006).
In the above-described embodiments, only enumerate " Chinese character " " hiragana ", still, also be applicable to any classification of classification character kinds such as " katakana " " English " " numeral " " symbol " " foreign language " in addition as the character kind.In addition, except that the character kind, also can use the affiliated product speech of this morpheme.
In addition, in the above-described embodiments, in order to specify the range of choice starting point or to switch preference pattern, and press " 1 ", " # " respectively, still,, then can distribute key arbitrarily so long as distributed the key of equivalent.
Moreover, in present embodiment, the centre coordinate of the cursor of cross mark is used as selection information, still,, then can be other information so long as give method with effect same.For example, two parantheses can be presented on the display, use the rectangular coordinates of the centre coordinate or the two parantheses of use of this parantheses.In addition, in the selection of Japanese, though be illustrated writing across the page,, equally also go for perpendicular writing.
In addition, the foregoing description can be distinguished realization separately, also can realize with the form that optionally adopts all or part of.
Claims (22)
1. portable terminal device is to have:
The image pickup section of photographic images,
Extract in the image character identifying object character row the character row extracting part,
The character recognition portion of the character in the recognition image,
Rotate described image with the image processing part that revises and
Be used to show the portable information terminal of the image displaying part of the image that becomes identifying object, it is characterized in that,
Show the angle display that shows shooting angle suitable for character recognition is handled quantitatively, the photographic images that is judged to be proper angle is carried out character recognition.
2. portable terminal device is to have:
From taken rest image, extract character identifying object character row the character row extracting part,
Discern the character in this image character recognition portion,
Rotate this image with the image processing part that revises and
Be used to show the portable information terminal of the image displaying part of the image that becomes identifying object, it is characterized in that,
Implement following the processing:
The rotation processing of the intact rest image of described shooting,
The extraction processing of character recognition object candidates character string,
From the character recognition object candidates character string of described extraction, select the selection of desirable identifying object character string handle and
Character recognition to described selecteed identifying object character string is handled.
3. portable terminal device according to claim 1 and 2 is characterized in that,
According to the asperratio of the boundary rectangle of the character row of described extraction, judge automatically and write across the page or perpendicular writing, and switch recognition mode.
4. portable terminal device according to claim 1 is characterized in that,
The direction of the display frame of the portable information terminal during according to image taking is judged automatically and is write across the page or perpendicular writing, and switches recognition mode.
5. portable terminal device according to claim 2 is characterized in that,
Follow the tracks of the inclination of the character string in the image of described shooting, automatically generate and show the boundary rectangle of identifying object candidate character strings.
6. portable terminal device is characterized in that having:
The image pickup section of photographic images,
Extract in this image character identifying object character row the character row extracting part,
Discern the character in this image character recognition portion,
Retrieval candidate generating unit, the picture that generates the retrieval candidate according to the output of described character recognition portion show described retrieval candidate generating unit output image displaying part and
Be used to select the user interface part of the retrieval candidate that described picture shows.
7. portable terminal device according to claim 6 is characterized in that,
Described retrieval candidate generating unit for the output of character recognition portion, uses plain analysis of voice to generate the retrieval candidate.
8. portable terminal device according to claim 6 is characterized in that,
Described retrieval candidate generating unit, the corresponding character class of each character code in the character string of distribution and character identification result, the character code that character class is identical is as the output of 1 retrieval candidate.
9. portable terminal device according to claim 6 is characterized in that,
Described retrieval candidate generating unit is used for the plain result who analyzes of voice of the output of character recognition and the two the information as a result after having distributed the corresponding character class of each character code with character identification result, output retrieval candidate.
10. portable terminal device according to claim 6 is characterized in that,
Described retrieval candidate generating unit, the information of each character code in the character string of use character identification result and any one among the geometry information generate the retrieval candidate.
11. portable terminal device according to claim 6 is characterized in that,
Described retrieval candidate generating unit uses plain analysis of voice to generate the retrieval candidate for the output of character recognition, in conjunction with the cursor position, generates and is used to select morpheme and has made up any 1 selection area coordinate among the compound of morpheme.
12. the character identifying method in the portable terminal device, the character of the character recognition object character row that identification is extracted from the image of taking is characterized in that, may further comprise the steps:
The step of photographic images,
Extract in the described image character identifying object character row step,
The image that rotates described shooting with revise the step that tilts,
Show the image become identifying object step,
Shooting angle suitable for character recognition is handled be shown to quantitatively the portable information terminal user step,
According to the shooting angle of described demonstration judge proper angle step and
The step of carrying out character recognition for the image of taking with the proper angle of described judgement.
13. the character identifying method in the portable terminal device, the character of the character recognition object character row that identification is extracted from the image of having taken is characterized in that, may further comprise the steps:
Extract in the described image character identifying object character row step,
Rotate described image with the step that revises,
Show the image become identifying object step,
Carry out the rotation processing of the intact image of described shooting step,
When carrying out described rotation processing, extract character identifying object candidate character strings step,
From the character recognition object candidates character string of described extraction, select the identifying object character string step and
The step of carrying out character recognition for described selecteed identifying object character string.
14. the character identifying method according in claim 12 or the 13 described portable terminal devices is characterized in that,
Carry out further comprising in the step of described character recognition:
According to the asperratio of the boundary rectangle of the character row that is extracted, automatically judge and write across the page or the perpendicular step of writing;
Switch the step of recognition mode based on described result of determination.
15. the character identifying method in the portable terminal device according to claim 12 is characterized in that,
The step of carrying out described character recognition further comprises:
The direction of the display frame of the portable information terminal during according to image taking is automatically judged and is write across the page or the perpendicular step of writing;
Switch the step of recognition mode based on described result of determination.
16. the character identifying method in the portable terminal device according to claim 13 is characterized in that,
From described character recognition object candidates character string, further comprise in the step of selection identifying object character string:
Track up finishes the inclination of the character string in the image, automatically generates and show the step of the boundary rectangle of identifying object candidate character strings.
17. the searching object word select selection method in the portable terminal device may further comprise the steps:
The step of photographic images,
Extract in the described image character identifying object character row step,
Discern the character in the described image step,
Use the step of the inside and outside data of being stored of result retrieval portable terminal device after the described character recognition, it is characterized in that, also comprise:
Based on the result after the described character recognition generate the retrieval candidate step,
Picture show described retrieval candidate step,
Select the step of the retrieval candidate of described picture demonstration.
18. the searching object word select selection method in the portable terminal device according to claim 17 is characterized in that,
Generate in the step of described retrieval candidate and further comprise:, use the plain step that generates the retrieval candidate of analyzing of voice for the result after the described character recognition.
19. the searching object word select selection method in the portable terminal device according to claim 17 is characterized in that,
The step that generates described retrieval candidate further comprises:
The step of the corresponding character class of each character code in the character string of the result after distribution and the described character recognition,
The character code that described character class is identical is as the step of 1 retrieval candidate output.
20. the searching object word select selection method in the portable terminal device according to claim 17 is characterized in that,
The step that generates described retrieval candidate is: use for the result after the described character recognition and carry out the plain result who analyzes of voice and distributed the two information as a result with the corresponding character class of each character code of character identification result, export the retrieval candidate.
21. the searching object word select selection method in the portable terminal device according to claim 17 is characterized in that,
The step that generates described retrieval candidate is: use the information of each character code in the character string of the result after the described character recognition and any one information among the geometry information, generate the retrieval candidate.
22. the searching object word select selection method in the portable terminal device according to claim 17 is characterized in that,
The step that generates described retrieval candidate further comprises:
For the result after the described character recognition, use the plain step that generates the retrieval candidate of analyzing of voice;
In conjunction with the position of cursor, generate and to be used to selecting morpheme and to have made up any 1 selection area coordinate among the compound of morpheme.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2003379288 | 2003-11-10 | ||
| JP2003379288A JP4443194B2 (en) | 2003-11-10 | 2003-11-10 | Processing object selection method in portable terminal character recognition and portable terminal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1617163A true CN1617163A (en) | 2005-05-18 |
| CN1292377C CN1292377C (en) | 2006-12-27 |
Family
ID=34689385
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004100889727A Expired - Fee Related CN1292377C (en) | 2003-11-10 | 2004-11-09 | Method for selecting treating object in character identification of portable terminal and portable terminal |
Country Status (4)
| Country | Link |
|---|---|
| JP (1) | JP4443194B2 (en) |
| KR (1) | KR100615058B1 (en) |
| CN (1) | CN1292377C (en) |
| TW (1) | TWI294100B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101482924B (en) * | 2008-01-08 | 2012-01-04 | 华晶科技股份有限公司 | Automatic identifying and correcting method for business card display angle |
| CN101674414B (en) * | 2005-09-09 | 2012-04-11 | 佳能株式会社 | Image pickup apparatus |
| CN103150088A (en) * | 2011-08-31 | 2013-06-12 | 三星电子株式会社 | Schedule managing method and apparatus |
| CN104461424A (en) * | 2014-12-01 | 2015-03-25 | 上海斐讯数据通信技术有限公司 | System and method for displaying rotary character strings in cells |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100754656B1 (en) * | 2005-06-20 | 2007-09-03 | 삼성전자주식회사 | Method and system for providing user with image related information and mobile communication system |
| JP4844142B2 (en) * | 2006-02-06 | 2011-12-28 | セイコーエプソン株式会社 | Printer |
| KR100641791B1 (en) | 2006-02-14 | 2006-11-02 | (주)올라웍스 | Tagging Method and System for Digital Data |
| US8208725B2 (en) | 2007-06-21 | 2012-06-26 | Sharp Laboratories Of America, Inc. | Methods and systems for identifying text orientation in a digital image |
| US8144989B2 (en) | 2007-06-21 | 2012-03-27 | Sharp Laboratories Of America, Inc. | Methods and systems for identifying text orientation in a digital image |
| JP2012008733A (en) * | 2010-06-23 | 2012-01-12 | King Jim Co Ltd | Card information management device |
| CN103377371A (en) * | 2012-04-25 | 2013-10-30 | 佳能株式会社 | Method and system for improving recognition features and optical character recognition system |
| JP5940615B2 (en) * | 2014-09-09 | 2016-06-29 | 株式会社アイエスピー | Skew logic character recognition method, program, and portable terminal device for portable terminal device |
| JP6371662B2 (en) * | 2014-10-07 | 2018-08-08 | 富士通フロンテック株式会社 | Character recognition support device, character recognition support program, and character recognition support method |
| KR101712391B1 (en) | 2015-06-22 | 2017-03-07 | 한국표준과학연구원 | In-situ graph analysis application for smart-phone |
| CN106325522B (en) * | 2016-09-05 | 2019-03-29 | 广东小天才科技有限公司 | A kind of method and apparatus that electric terminal adjusts cursor size |
| KR102391068B1 (en) * | 2020-07-24 | 2022-04-28 | 엄춘호 | Document recognition system and method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3281469B2 (en) * | 1993-11-18 | 2002-05-13 | 株式会社リコー | Document image inclination detecting method and apparatus |
| JPH11250179A (en) * | 1998-02-27 | 1999-09-17 | Matsushita Joho System Kk | Character reocognition device and its method |
-
2003
- 2003-11-10 JP JP2003379288A patent/JP4443194B2/en not_active Expired - Fee Related
-
2004
- 2004-10-04 TW TW093129992A patent/TWI294100B/en not_active IP Right Cessation
- 2004-11-04 KR KR1020040089371A patent/KR100615058B1/en not_active IP Right Cessation
- 2004-11-09 CN CNB2004100889727A patent/CN1292377C/en not_active Expired - Fee Related
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101674414B (en) * | 2005-09-09 | 2012-04-11 | 佳能株式会社 | Image pickup apparatus |
| CN101482924B (en) * | 2008-01-08 | 2012-01-04 | 华晶科技股份有限公司 | Automatic identifying and correcting method for business card display angle |
| CN103150088A (en) * | 2011-08-31 | 2013-06-12 | 三星电子株式会社 | Schedule managing method and apparatus |
| CN103150088B (en) * | 2011-08-31 | 2018-03-16 | 三星电子株式会社 | Agenda managing method and equipment |
| CN104461424A (en) * | 2014-12-01 | 2015-03-25 | 上海斐讯数据通信技术有限公司 | System and method for displaying rotary character strings in cells |
| CN104461424B (en) * | 2014-12-01 | 2017-11-03 | 上海斐讯数据通信技术有限公司 | A kind of system and method that rotation character string is shown in cell |
Also Published As
| Publication number | Publication date |
|---|---|
| KR20050045832A (en) | 2005-05-17 |
| CN1292377C (en) | 2006-12-27 |
| KR100615058B1 (en) | 2006-08-22 |
| TW200516509A (en) | 2005-05-16 |
| TWI294100B (en) | 2008-03-01 |
| JP2005141603A (en) | 2005-06-02 |
| JP4443194B2 (en) | 2010-03-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1292377C (en) | Method for selecting treating object in character identification of portable terminal and portable terminal | |
| CN101667251B (en) | OCR recognition method and device with auxiliary positioning function | |
| US10248878B2 (en) | Character input method and system as well as electronic device and keyboard thereof | |
| JP6138305B2 (en) | Camera OCR using context information | |
| CN1269014C (en) | Character input device | |
| KR101220709B1 (en) | Search apparatus and method for document mixing hangeul and chinese characters using electronic dictionary | |
| CN1839396A (en) | Document scanner | |
| CN100336375C (en) | Portable terminal device and character input method | |
| JP2014102669A (en) | Information processor, information processing method and program | |
| JPWO2007004519A1 (en) | Search system and search method | |
| CN101076166A (en) | Device having display buttons and display method and medium for the device | |
| JP2005346707A (en) | Low-resolution ocr for document acquired by camera | |
| CN1940941A (en) | Image analysis apparatus and image analysis program storage medium | |
| KR100759165B1 (en) | Portable terminal and character reading method using a portable terminal | |
| KR20210086836A (en) | Image data processing method for searching images by text | |
| CN110806407A (en) | Labview-based two-dimensional material scanning and vision processing system and method | |
| EP2428884A2 (en) | Method, software, and apparatus for displaying data objects | |
| CN1250205A (en) | Document image processing apparatus and its method and recording medium with all program | |
| CN1234063C (en) | Handwriting characters inputting supporter and its method | |
| CN110795918B (en) | Method, device and equipment for determining reading position | |
| CN1317664C (en) | Confused stroke order library establishing method and on-line hand-writing Chinese character identifying and evaluating system | |
| CN1606030A (en) | Electronic photography translation paraphrasing method and apparatus | |
| CN1107280C (en) | Chinese and English table recognition system and method | |
| JP5325870B2 (en) | Character string output device, character recognition system, program, and character string output method | |
| CN113407757B (en) | Image retrieval method and device based on computer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20061227 Termination date: 20131109 |