JPH08147416A - Information processing method - Google Patents

Abstract

PURPOSE: To speedily output a candidate word when the candidate word matches an index word in a dictionary and to easily output all matching index words by outputting a 1st candidate word by a matching processing using the word dictionary, and then continuously matching other character strings with the word dictionary so as to detect a 2nd candidate word. CONSTITUTION: An image is read by a scanner 1 and converted into a binary signal, which is stored in an image buffer in a RAM 11. A character recognizing processing which matches the image data with the dictionary for character recognition and codes the data is performed and the result is stored in the RAM 11. All candidate words as candidates for a word in all combinations obtained by replacing the recognition result with 1st candidates and 2nd candidates of respective character patterns are compared with the index words of the dictionary stored in a ROM 12 in order. Matching words are all stored in the RAM 11 and called in sequence when necessary. When a 1st matching word is not a desired word, other words can be called through easy operation.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing method for recognizing a character and searching a dictionary or the like using the recognized character string.

[0002]

2. Description of the Related Art Conventionally, in this type of apparatus, a character string that recognizes an input character pattern is used to search a dictionary or the like and output the most suitable heading word from the dictionary heading words. is there.

[0003]

However, in the above-mentioned conventional example, since the output headword is limited to one word, even if a plurality of words are generated by exchanging characters and the dictionary search is repeated, I had to somehow select any one of the dictionary headwords.

There are various methods for character recognition, such as a pattern matching method and a structure analysis method. However, the wider the character font to be recognized, the more difficult it is to recognize the character, and the candidate character of each character is unique. The things that are not decided tend to happen. In such a case, multiple words can become candidate words for the input character pattern by replacing the candidate characters. However, if the conventional method of selecting any one word is used, all the words are replaced. Also, there is a drawback in that the output cannot be performed until the search is completed and the most suitable word is selected, and the operator has to wait for that time. Further, even if two or more words out of the plurality of candidate words are fully matched with the dictionary word in the dictionary, one of them has to be determined, so that a correct word may not be output.

The present invention has been made in view of the above-mentioned drawbacks of the conventional example, and an object thereof is to promptly output and match when a plurality of candidate words match with a dictionary entry word. It is to easily output all the entered headwords.

[0006]

[Means for Solving the Problems] In order to solve the above problems,
The present invention recognizes input image information to extract a plurality of recognition candidate characters, matches a character string obtained from the combination of the recognition candidate characters with a word dictionary, and matches candidate words by matching with the word dictionary. An information processing method for outputting, the first of which is performed by matching with the word dictionary.
Provided is an information processing method which, after outputting the second candidate word, continues matching other character strings with the word dictionary to detect the second candidate word.

In order to solve the above-mentioned problems, the present invention preferably notifies that the second candidate word has been detected when the second candidate word has been detected by the matching processing.

In order to solve the above problems, the present invention preferably sequentially stores candidate words detected by matching with the word dictionary.

In order to solve the above problems, the present invention preferably obtains the character string by indexing a word region from the input image and combining candidate characters in the indexed word region.

In order to solve the above problems, the present invention preferably sequentially outputs the stored candidate words in response to the input of the next candidate output instruction.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram showing the configuration of an information processing apparatus according to this embodiment. Reference numeral 1 is a scanner for optically reading a document image and inputting an input character pattern.
It may be a large one that can read a document of a size or the like, or a handy type that is easy to operate. Reference numeral 2 is a binarization circuit that converts an analog signal output from the scanner into a binary digital signal, and 10 is a CPU that performs character recognition processing, dictionary search processing, and other arithmetic processing.
The processing is executed according to the control program stored in the ROM 12. Reference numeral 11 is a RAM having an image buffer for storing the image data output from the binarization circuit 2 and each work buffer, and 12 is a program for each process such as a control program of a flowchart described later and various patterns such as standard patterns used for character recognition. ROM that stores a dictionary of words or phrases that match data and recognition results, 13
Is a driver for driving a display such as the LCD 14, and the LCD 14 is a liquid crystal display for displaying output data. The display may be an LCD, a CRT or the like as long as it can display the recognition result. Reference numeral 15 denotes a keyboard or touch keys on the display unit 14 to input various instructions.

FIG. 2 is a diagram showing an example of a character string to be read, 21 is an area to be recognized, which is read by the scanner 1 or designated on the display after reading, 22 is an area 21 shows a range of words cut out by performing image processing on 21 as a preprocessing of the recognition processing.
The extraction of the word is performed by taking a histogram of the image in the vertical and horizontal directions, and regarding a character string portion sandwiched by a large space portion as a word.

FIG. 3 shows the recognition result of each character pattern of the cut word, which is stored in the RAM 11 in the form of a character code. The first row shows the maximum similarity between character patterns.
The lower rank shows the second rank candidate whose similarity is the second best.

FIG. 4 shows candidates of all combinations of words obtained by exchanging the first-ranked candidates and the second-ranked candidates of each character pattern based on the recognition result of FIG.
The headwords in the dictionary stored in the OM 12 are sequentially compared.

FIG. 5 shows an example of words in the recognition result display frame displayed on the LCD 14. Reference numeral 51 denotes a headword word display section registered in the dictionary of matched candidate words, and 52.
Is a part of speech stored in correspondence with the heading 51 in the dictionary,
A meaning display portion for displaying classification, meaning, etc., 53 is a pronunciation valid display 54, a idiom valid display 55 is a next candidate valid display (5
For each of the valid indications 3 to 55, there is a key on the keyboard 15 capable of giving an instruction corresponding to each valid indication.
It indicates that each key is valid when 3 to 55 are lit (displayed). These effective display 53
If ~ 55 is not performed, the corresponding key is invalidated or a process of notifying that there is no information to be displayed even if the key is pressed is performed).

The display transitions of (a) to (c) in FIG. 5 will be described with reference to the flowcharts of FIGS. 6 and 7.

First, in step S1 of FIG.
The command input by the user is received, and the command indicated by each key is determined in step S2. If it is determined in S2 that the read key has been pressed, the read recognition process of step S3 is performed. If it is determined in S2 that the next candidate key is pressed, the next candidate process is performed in step S4. If it is determined in step S2 that another key is pressed, other processing is performed in step S5. A display process of displaying the data set in the display buffer in each process of S3 to S5 on the LCD 14 in step S6 is performed.

FIG. 7 is a flow chart showing the read recognition processing in step S3 in detail. In step S11, the image is read by the scanner 1, binarized by the binarization circuit 2, and the image is stored in the image buffer in the RAM 11. The image input process is performed. A character recognition process for matching and encoding with a character recognition dictionary is performed and stored in the RAM 11 in the format as shown in FIG. In step S13, a word cutout process is performed to cut out a word from the information of each recognized character. The order of steps S13 and S12 may be reversed. In step S14, a word dictionary search process is performed using the character code of the recognition result of the cut word portion. If it is determined in step S15 that the dictionary is present, the process proceeds to step S16. Here, it is checked whether or not the word determined to be present in step S15 is the first word that matches the dictionary by checking whether or not the matching word is already displayed on the display unit 14. If the candidate word is not displayed and it is determined that it is the first word, the process proceeds to step S17, information about the word is read from the dictionary of the ROM 12, and the headword and its meaning are displayed on the display screen of the display unit 14.

In the example of FIG. 4, 41 "bat" matches the dictionary first, so this word is first displayed on the display 14
To be displayed. FIG. 5 (a) shows this state. Next, in step S19, the headword is stored in the RAM 11
It stores in the candidate word storage area in. And step S2
Proceeding to step 0, all combinations of candidate characters stored in the RAM 11 are created, and it is checked whether or not the replacement of characters has been completed. If not yet completed, step S21
Go to and replace the character strings. After the hat of 41, 42 "hat" is obtained by replacing the first candidate character of the first character with the second candidate. Then again step S14
Search the dictionary in ROM12 with this hat41
Is in the dictionary, the process proceeds from step S15 to step S16. Here, hat41 is the second matched word, and the candidate word has already been displayed on the display unit 14, so the process proceeds to step S18, and the next candidate valid display 55 is lit on the display unit 14. This is the state shown in FIG.

Next, a process of storing the headword in the candidate word storage area in the RAM 11 is performed in step S19. Thereafter, the characters are sequentially replaced in step S21 until it is determined in step S20 that there are no replacement characters, and the dictionary is repeatedly searched. If it is determined in step S20 that there are no replacement characters, the read recognition process ends.

Next, returning to the flow chart of FIG. 6 again, if the above candidate words are stored in the RAM 11 and it is determined in step S2 that the next candidate key has been pressed with the next candidate valid display 55 lit, step S4. In the next candidate process, the second matching "hat" with the dictionary is called from the candidate word storage area of the RAM 11, and the information of that word is RO.
The heading and the meaning are read from the dictionary of M12 and set in the display buffer, and the display process is performed in step S6. This is the state shown in FIG.

In this way, the dictionary entry word "hat" and its meaning for the actually read word "hat" can be obtained.

FIG. 8 shows an example of the display method when the second candidate is searched in step 18 instead of the display of FIG. 5 (b) described above. The first and second dictionary headwords matched in step S15 are collectively displayed on the display screen of the display 14, and are then moved as indicated by arrows (a) and (b). The desired headword is designated, and the contents of the dictionary are actually displayed in response to the depression of the dictionary content display instruction key of the keyboard 15. In 8- (a), "bat" is instructed.
In (b), "hat" is designated. When a selection key (not shown) for instructing the display of the dictionary contents is pressed here, the screen shown in FIG. 5C can be displayed.

In the description of step S15 in FIG. 7, the matching process with the dictionary is a full match whether the dictionary is present or absent. It may be a match.

[0025]

As described above, according to the present invention, each of a plurality of character strings obtained as a result of character recognition is matched with a dictionary, and all words successfully matched with the dictionary headword are stored in a RAM. Since it can be called sequentially as needed, even if the first matched word or the word judged to be the most suitable is not the desired word, it will be notified that there are other matched words and other words You can call a word with a simple operation.

Further, even if matching with all word dictionaries created by exchanging the results of character recognition is not completed, it is possible to perform display processing of that word at the time of the first match, so the operator The waiting time can be shortened.

[Brief description of drawings]

FIG. 1 is a block diagram of the configuration of an information processing apparatus that implements the present invention.

FIG. 2 is a diagram showing a character string to be read.

FIG. 3 is a diagram showing a recognition result.

FIG. 4 is a diagram showing word candidates obtained by a replacement process.

FIG. 5 is a diagram showing an LCD displaying words.

FIG. 6 is a flowchart showing overall control.

FIG. 7 is a flowchart showing a reading recognition process.

FIG. 8 is a diagram showing a display example for selecting a word.

Claims (5)

[Claims] 1. A plurality of recognition candidate characters are extracted by recognizing input image information, a character string obtained from a combination of the recognition candidate characters is matched with a word dictionary, and a matching candidate is obtained by matching with the word dictionary. An information processing method for outputting a word, comprising: outputting a first candidate word by matching with the word dictionary, and then matching another character string with the word dictionary to detect a second candidate word. An information processing method characterized by continuing. 2. The information processing method according to claim 1, wherein when the second candidate word is detected by the matching process, the fact that the second candidate word is detected is notified. . 3. The information processing method according to claim 1, wherein candidate words detected by matching with the word dictionary are sequentially stored. 4. The information processing method according to claim 1, wherein a word area is indexed from the input image, and the character string is obtained by combining candidate characters in the indexed word area. 5. The information processing method according to claim 3, wherein the stored candidate words are sequentially output in response to an input of a next candidate output instruction.