JP3233803B2 - Hard-to-read kanji search device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an obfuscated kanji retrieval apparatus for retrieving kanji (obfuscated kanji) whose reading is unknown in kanji input by a word processor or the like.

[0002]

2. Description of the Related Art Conventionally, in a document creation process using a word processor or the like, when searching for kanji whose reading is unknown, that is, obfuscated kanji, (1) the total number of strokes (2) the number of radicals and the number of strokes other than radicals Kanji search is performed from the Kanji code table by the method. That is, in the related art, retrieval of obfuscated kanji is performed using a method based on the number of strokes.

[0003]

However, the search for obfuscated kanji by the stroke number method has the following problems. (1) Most of the difficult-to-read kanji have many strokes, and it takes time to grasp the exact stroke count. Generally, the correct answer rate for 10 or more strokes is particularly low.

[0004] (2) Since difficult-to-read kanji are generally other than kanji for use, there are many cases where old characters are used and it is difficult to accurately determine the number of strokes. (3) The number of strokes of kanji is mostly 5 to 20 strokes, and can be divided into only a dozen or so. For this reason, the number of kanji existing in the same stroke count is large, and if the stroke count is wrong, it takes time to search.

(4) Kanji cultural sphere (China, Korea, Japan, N
IES etc.), because the fonts differ in detail,
In many cases, the number of strokes is wrong. In recent years, exchange has been rapidly deepening due to the rapid growth of economic power in the kanji cultural sphere. In such an environment, mutual proper nouns (place names, person names, etc.)
Because it cannot be translated, it will be expressed as it is,
Many of these types of kanji are difficult to read. The present invention has been made in view of the above circumstances, and has as its object to provide an obfuscated kanji search device capable of searching for an obfuscated kanji in a short time.

[0006]

According to the present invention, there is provided an apparatus for retrieving obfuscated kanji characters, wherein a radical of each kanji and the first N of a part other than the radical are used.
Image (N is a desired natural number) "\" to write out lines of,
Classified into four types of "-", "|", "/"
Classification of the total stroke count dictionary and each kanji by radical
And the write-out line of the first N drawing except the radical
"＼", "-", "|", "/"
A radical type dictionary that is classified and stored, and a total stroke count mode
Search mode for selecting the search mode
Selection means and "＼", "-",
A pattern for inputting four types of patterns, "|" and "/"
Enter key and total search mode
Approximate stroke count input when a stroke is selected
Means and the search mode selection means set the total number of strokes mode.
If selected, enter it using the pattern input keys described above.
Of the first N strokes of the radical and non-radical part of the search kanji
The turn and the approximate stroke number input by the approximate stroke number input means
Search the total stroke count dictionary based on the stroke stroke count and select candidate Kanji
Means for outputting the total number of strokes pattern, and the above search mode
When the radical mode is selected by the selection means,
Display the input screen and enter the search kanji radical entered on this screen.
And the search kanji input with the pattern input key
According to the pattern of the first N pictures of the part other than the radical,
A radical pattern search that searches another dictionary and outputs candidate kanji
Cable means .

[0007]

In the above-mentioned storage means, the first writing line of each of the first N strokes (N is a desired natural number) of each kanji radical and the portion other than the radical is set to "@", "-", "| , "/" Are stored in four patterns. Then, when retrieving the obfuscated kanji, the radicals of each kanji to be retrieved and the first writing line of each of the first N strokes other than the radicals are input in the patterns classified into the above four types. Then, a kanji search is performed by referring to the storage means based on the input pattern, and the search result is displayed on the screen as candidate kanji.

As described above, the first writing line of each of the first N strokes of each kanji radical and the portion other than the radical is divided into four types of patterns and stored in the storage means. Thus, a desired kanji can be searched in a short time by inputting a combination of four types of patterns without knowing how to read the kanji.

[0009]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a functional configuration of an obfuscated kanji search device in a kanji input device such as a word processor.

In FIG. 1, reference numeral 11 denotes a search mode selection device for selecting one of two search modes, that is, a total stroke number mode or a radical mode. Reference numeral 12 denotes a pattern input device, which sets the first writing lines of the first N strokes (N is a desired natural number) of the radicals and non-radical portions of each kanji to be searched by "@", "-", " | And "/" are input. In this case, in the case of the total stroke number mode, the pattern of the N picture including the radical is input. In the case of the radical mode, the radical is designated, and then the pattern of the first N picture other than the radical is inputted. I do. The above four patterns are
"＼", Which JIS keyboards generally have,
The "-", "|", and "/" keys are used, but they may be separately provided as pattern input keys.

Reference numeral 13 denotes a rough stroke number input device, which is used to change the rough stroke number of the search character into, for example, a small stroke number (10 or less) and a medium stroke number (8 to 1).
6 strokes) and the number of large strokes (14 strokes or more). In classifying the number of strokes, it is better to overlap several strokes. Since the approximate stroke number input device 13 is used as an auxiliary device, it can be omitted.

Reference numeral 14 denotes a pattern search device for performing a kanji search by a pattern with reference to a total stroke count dictionary 15 and a radical type dictionary 16. In the total stroke count dictionary 15, Chinese characters including radicals are classified and stored in advance in four types of patterns, "@", "-", "|", and "/". In the radical type dictionary 16, kanjis are classified by radical, and parts other than radicals are classified into four types of "@", "-", "|", and "/". It is remembered. The pattern dictionary creating apparatus for creating the total stroke count type dictionary 15 and the radical type dictionary 16 will be described later in detail.

A candidate kanji output device 17 selects a kanji having a high correct answer probability from the kanji searched by the pattern search device 14. The candidate kanji selected by the candidate kanji output device 17 is sent to the extraction result determination input device 18 and displayed on the screen. The operator inputs the success or failure of the candidate kanji displayed on the screen. Based on the input, the determination unit 19 determines whether the candidate kanji is correct or incorrect. If the candidate kanji is incorrect, the pattern re-input device 20 re-inputs the pattern or the pattern additional input device 21 An additional pattern is input, and the pattern search device 14
Send to The pattern search device 14 executes the pattern search again based on the pattern. When the correct candidate kanji is extracted, the kanji search ends.

FIG. 2 is a block diagram showing a schematic configuration of a pattern dictionary creating device 30 for creating the total stroke count dictionary 15 and the radical type dictionary 16. Pattern dictionary creation device 3
Numeral 0 comprises a kanji direction encoding device 31, a total stroke count type pattern extracting device 32, a radical type pattern extracting device 33, and a pattern classifying device. The kanji direction coding device 31 converts the writing method (corresponding to one stroke) for each target kanji in accordance with the stroke order, as follows: (1) “＼”, (2)
It is patterned into four types of “-”, (3) “|”, and (4) “/”. That is, each target kanji is divided into the above four types for each stroke in accordance with the stroke order, that is, (1) "diagonal line at lower right", (2)
The pattern is formed into “horizontal line”, (3) “vertical line”, and (4) “lower left diagonal line”.

For example, in the case of the Chinese character "width", the "width" of the radical is "vertical", "horizontal", "vertical" in the stroke order, that is,
The pattern is "|", "-", "|". In the same way, for parts other than the radical, "-",
"|", "-", "-", "|",... Other kanji are similarly coded using four types of patterns.

The total stroke number type pattern extracting device 32 extracts the first N strokes including the radicals of the kanji coded by the kanji direction coding unit 31, for example, a pattern of five strokes. For example, in the case of the kanji “width”, a combination of patterns “|”, “−”, “|”, “—”, “|” is used.

The radical type pattern extracting device 33 extracts the first N strokes, for example, five strokes of the kanji characters coded by the kanji direction encoding device 31 except for the radicals. . For example, in the case of the kanji character "width", a combination of patterns "-", "|", "-", "-", "|"

The pattern classifying device 34 classifies the patterns extracted by the total stroke number type pattern extracting device 32 to create the total stroke number type dictionary 15 and extracts the radical stroke type pattern extracting device 33. The patterns are classified according to radicals, and a radical-specific dictionary 16 is created.

The obfuscated kanji search device shown in FIG. 1 performs the obfuscated kanji search process using the total stroke count type dictionary 15 and the radical type dictionary 16 created as described above. Hereinafter, a search operation of the obfuscated kanji search device will be described with reference to a flowchart shown in FIG.

First, an obfuscated kanji search mode is designated on the document creation screen by using the function of the search mode selection device 11 (step A1). Search mode selection device 11
When the obfuscated kanji search mode is designated, next, a screen for selecting the "total stroke number mode" or the "radical mode" is displayed (step A2). Now, assuming that the "total stroke number mode" is designated here, the pattern input device 12 displays a pattern input screen for obfuscated kanji (step A3). On this display screen, the pattern based on the total number of strokes of the obfuscated kanji that you want to search is "＼", "-", "|"
Using the “/” key, for example, the first five strokes are input. For example, in the case of Chinese characters “width”, “ido”, and “hui”, “width”: “|”, “—”, “|”, “—”, “|” “ido”: “/”, “ "＼", "/", "-", "-""Hui": Input as "/", "/", "|", "|", "|".

After the pattern is input as described above, the approximate stroke number of the search kanji is input using the function of the approximate stroke number input device 13. For example, the number of small strokes (10 strokes or less), the number of medium strokes (8 to 16 strokes), and the number of large strokes (14 strokes or more) are input separately.

When the "total stroke count mode" is designated as described above, the pattern retrieval device 14 retrieves obfuscated kanji characters using the total stroke count dictionary 15 within the designated stroke count (step A5). In this case, the obfuscated kanji can be searched without specifying the number of strokes, but it is more efficient to specify the number of strokes.

The pattern search device 14 searches the total stroke count dictionary 15 based on the input pattern, and outputs the search result to the candidate kanji output device 17. The candidate kanji output device 17
A kanji having the highest probability of a correct answer is selected from the searched kanji and displayed at a predetermined position on the screen (step A6).
If the displayed candidate kanji is the desired kanji, the search process is terminated by performing a confirmation operation or the like. If not, the next candidate kanji is changed by operating the next candidate key. Display. When a desired kanji cannot be obtained as a candidate kanji, the pattern input may be incorrect or the number of input strokes of the pattern may be too small to extract the desired kanji. In this case, a new pattern is re-input by using the function of the pattern re-input device 20, or a pattern of the sixth and subsequent strokes is additionally input by using the function of the pattern additional input device 21. And
The search process is executed again by the pattern search device 14. When the desired kanji is obtained by re-executing the pattern search, the search process is terminated.

When the "radical mode" is designated in step A2, the search mode selection device 11 displays a radical input screen and allows the user to input the radical of the search kanji (step A7). For example, in the case of the kanji “width”, “dou”, and “hui”, enter the radicals of “width” → “width”, “ido” → “mu”, “hui” → “hi”. When the radical is input, a pattern input screen for obfuscated Chinese characters is displayed by the pattern input device 12 (step A8). In this display screen, the pattern of the part other than the radical of the obfuscated kanji to be searched is indicated by “＼”,
Using the "-", "|", and "/" keys, the first five strokes are input in the same manner as in step A. For example,
For the kanji characters "width", "ido", and "hui", "width": "-", "|", "-", "-", "|""ido":"/","-" , “—”, “/”, “＼” “Hu”: Input as “|”, “|”, “|”, “/”, “/”. In the case of the “radical mode”, it is not necessary to input the total number of strokes.

When the "radical mode" is designated as described above, the pattern search device 14 searches the radical type dictionary 16 based on the input pattern, and outputs the search result to the candidate kanji output device. 17 (step A9). The candidate kanji output device 17 selects a kanji having the highest probability of correct answer from the searched kanji and displays it at a predetermined position on the screen (step A10). If the displayed candidate kanji is the desired kanji, the search process is terminated by performing a confirmation operation or the like. If not, the next candidate kanji is changed by operating the next candidate key. Display. If the desired kanji cannot be extracted, the pattern is re-input as in the case of the “total stroke number mode”, or the pattern after the sixth stroke is additionally input and the search process is executed again.

By inputting four types of patterns for N strokes in the stroke order as described above, it is possible to search for obfuscated Chinese characters in a short time. In general, even if you do not know how to read kanji, there are few people who make a mistake in the basic writing order. In addition, even if a mistake is made, there are only a few cases, and a correct answer can be easily obtained.

As shown in the above embodiment, when four types of patterns are input up to five strokes of obfuscated kanji, the combinations are 4 to the fifth power, that is, "4 ⁵ = 1024". It is possible to perform a structured search. For example, in the Kangxi dictionary, the number of characters is about 50,000, and even if characters with different fonts (simplified characters, frequent characters) are added, it is about 100,000 characters. Even if all the characters are converted into a database, the data amount is not so large. .

In the above-described embodiment, when a difficult-to-read kanji is searched, only the extracted kanji is displayed. However, the reading is stored in the total stroke count dictionary 15 and the radical type dictionary 16 together with the kanji. Alternatively, when a search is performed, the reading may be displayed together with the corresponding kanji.

The present invention is not limited to a word processor, but can be implemented in a computer having a kanji processing function in the same manner as in the above embodiment.

[0030]

As described in detail above, according to the present invention, the first writing line of each of the first N strokes of each kanji radical and the part other than the radical is divided into four types of patterns and stored in a dictionary. Since the dictionary is stored and searched based on the input patterns of N strokes, a desired kanji can be searched in a short time without knowing how to read the kanji.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an obfuscated kanji search device according to one embodiment of the present invention.

FIG. 2 is a block diagram showing a schematic configuration of a pattern dictionary creating apparatus for creating a dictionary in the embodiment.

FIG. 3 is a flowchart showing a processing operation of the embodiment.

[Explanation of symbols]

 11 Search Mode Selection Device 12 Pattern Input Device 13 Approximate Stroke Number Input Device 14 Pattern Search Device 15 Total Stroke Type Dictionary 16 Radical Type Dictionary 17 Candidate Kanji Output Device 18 Extraction Result Judgment Input Device 19 Judgment Unit 20 Pattern Reinput Device 21 Pattern Additional input device 30 Pattern dictionary creation device 31 Kanji direction coding device 32 Total stroke type pattern extraction device 33 Radical-specific pattern extraction device 34 Pattern classification device

Continuation of the front page (56) References JP-A-63-182769 (JP, A) JP-A-2-289100 (JP, A) JP-A-6-96266 (JP, A) JP-A-4-195671 (JP) JP-A-2-18661 (JP, A) JP-A-55-66017 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 17/21-17/26

Claims (1)

(57) [Claims] [Claim 1] "\" to write out lines of the first of the N image of a portion other than the radical and radical of each Chinese character (N is a desired natural number),
Classified into four types of "-", "|", "/"
The stored total stroke count dictionary and each kanji are classified by radical and the
The first N drawing lines are "＼", "-", "|",
Radicals classified and stored in four patterns of "/"
Select different dictionary and search mode of total strokes mode or radical mode
Search mode selection means and four characters “検 索”, “-”, “|”, “/”
The pattern input key for inputting the type of pattern and the total number of strokes mode are selected by the search mode selection means.
In this case, the rough stroke number input means for inputting the rough stroke number and the search mode selecting means select the total stroke number mode.
Search key entered using the pattern input key
The pattern of the first N strokes of the radical of the character and the part other than the radical
And the approximate stroke number input by the above approximate stroke number input means.
Based on the above total stroke count dictionary and output candidate Kanji
The radical mode is selected by the total stroke number pattern searching means and the search mode selecting means.
Is displayed, the radical input screen is displayed and the
Enter using the search kanji radical and the pattern input keys described above.
Pattern of the first N strokes of the part other than the radical of the searched kanji
A section that searches the radical dictionary and outputs candidate kanji
A difficult-to-read kanji search device comprising: a neck-specific pattern search means .