JPS63146187A - Character recognizing device - Google Patents

Abstract

PURPOSE:To easily find that a character is misrecognized as a character to which a character form is similar by changing display methods by respective character kinds and displaying recognition results so that the character kinds to which the characters belong are easily and visually understood. CONSTITUTION:A character kind decision part 3 decides a character kind corresponding to the 1st candidate of each recognition result and a recognition result display part 4 displays a character string which is the 1st candidate of the recognition result. At this time, alphanumeric characters, HIRAGANA (cursive form of Japanese syllabary) and KATAKANA (square form of Japanese syllabary) characters, and KANJI (Chinese character) are displayed while underscored with a dotted line, a full line, and a double full line respectively according to the decision result of the character kind. An operator finds misrecognition visually by using the displayed character string and character kind information added thereto. Thus, the character kind information added to each displayed character is used as a clue to find that a character is misrecognized as a character to which character form is similar, and the error is easily corrected.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a character recognition device that recognizes a character string included in a document input as image data and converts it into a character code string.

2. Description of the Related Art Character recognition devices that recognize characters written on paper and convert them into character codes that can be handled by computers have been widely developed.

However, most of these character recognition devices involve recognition errors, and a major challenge in developing character recognition devices is how to increase the recognition rate.

In early character recognition devices, for example, as seen in automatic postal code reading devices, character types and fonts to be handled,
By limiting the entry area, a recognition rate sufficient for practical use was ensured.

In recent years, these limitations are gradually being removed due to advances in character recognition technology. In addition to alphanumeric characters, hiragana and katakana are also available.

A Japanese character recognition device including kanji has been developed. As for fonts, fonts that can be used in both printed and handwritten fonts have been developed, and fonts that do not require limitation of the character area due to an automatic cutting function have also been developed.

FIG. 2 is a block diagram of a conventional character recognition device.

In the figure, reference numeral 11 denotes an image input section for inputting a document as image information. As the image input unit 11, for example, an image scanner is used. 12 is a character recognition unit.
The character string included in the document image input from the image input section 11 is recognized and converted into a character code string. For example, the J I 5-C-6226 code is used as the character code to be converted. Reference numeral 13 denotes a recognition result display section for displaying each character as a result of recognition by the character recognition section 12. With such a configuration, character strings included in a document input as an image can be recognized, and the results can be displayed for confirmation and error correction by the operator.

By the way, there is still no definitive solution to the problem of recognition errors, and the current situation is that recognition errors are detected and corrected by an operator visually.

In this case, in a character recognition device that can recognize characters of multiple character types, recognition errors are particularly likely to occur when characters of different character types with very similar shapes are recognized. is often not noticed by the operator.

For example, the English symbols “-()・ifun and minus)”
”, the long sound symbol “-”, the Chinese numeral “-”, etc., it is technically almost impossible to correctly recognize each other, and even if a character with an incorrect mental recognition result is displayed In conventional character recognition devices, the differences are not displayed very clearly, so operators often do not notice recognition errors.Overlooking recognition errors like the one described above can lead to people not using the character recognition device as a mere copying device. In some cases, it may not be such a problem. However, the character codes for each of these candidates are clearly different, which is a major problem when using the character recognition device as an input device for information processing by a computer.

Problems to be Solved by the Invention In conventional character recognition devices, when displaying the recognition results, the differences between different character types with very similar shapes were not very clear, so it was difficult to press the button or the operator. This type of recognition error was often overlooked when visually confirming the validity of the recognition result.

In view of the above-mentioned problems that conventional character recognition devices had, the present invention makes it possible to visually understand the character type to which a character belongs by changing the display method for each character type. It is an object of the present invention to provide a character recognition device that displays recognition results and can easily detect errors in recognition of characters of different character types with similar shapes.

Means for Solving the Problems In order to achieve the above object, the present invention provides an image input section for inputting a document as image information, and a recognition of character strings included in the document image input from the image input section. a character recognition unit that converts the character code into a character code string; a character type determination unit that determines to which character type each character code recognized by the character recognition unit belongs among a plurality of character types classified in advance; This constitutes a character recognition device that includes a recognition result display section that displays the character type to which the character belongs so that it can be visually and easily understood.

According to the above-described configuration, the present invention inputs a document as an image, recognizes a character string included in the input image, converts it into a character code string, and converts each character code of the recognition result into a plurality of pre-classified character codes. By determining which character type a character belongs to among the character types, and changing the display method for each character type according to the determination result, it is possible to display the character type to which the character belongs so that it can be easily understood visually. It is something.

Example An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a character recognition device according to an embodiment of the present invention. In the figure, reference numeral 1 denotes an image input unit for inputting a document as image information. As the image input section 1, for example, an image scanner is used. Reference numeral 2 denotes a character recognition section which recognizes character strings included in the document image input from the image input section 1 and converts them into character code strings. For example, the character code to be converted is TIS-C-6226.
Codes etc. are used. Reference numeral 3 denotes a character type determination unit which determines to which character type each character code recognized by the character recognition unit 2 belongs among a plurality of character types classified in advance. In this embodiment, three types of characters to be recognized are considered: alphanumeric characters, hiragana and katakana characters, and kanji characters. The three sets mentioned above have nothing in common with each other, and the character codes for the characters belonging to each set are assigned to a specific continuous area on the character code set, so the character can be read from any character code. Determining which of the above character types corresponds to can be done very easily by simple numerical comparison. 4
is a recognition result display section, which changes the display method for each character type based on the result of recognition by the character recognition section 2 according to the determination result by the character type determination section 3, so that it is easy to visually identify the character type to which the character belongs. It is displayed in such a way that it can be easily understood. In this embodiment, alphanumeric characters are underlined with dotted lines, hiragana and katakana with solid lines, and kanji with double solid lines.

The operation of the character recognition device of the present invention configured as described above will be explained below.

(1) The operator inputs multiple documents containing character strings to be recognized through the image input unit 1.

(2) The character recognition unit 2 recognizes character strings included in the input document image and converts them into character code strings.

Recognition results are obtained for each character up to the Nth candidate (N is a constant natural number).

(3) The character type determining unit 3 determines the corresponding character type for each first candidate of the recognition results.

(4) The recognition result display section 4 displays the first candidate character string of the recognition result. At this time, according to the character type determination result, alphanumeric characters are marked with dotted lines, hiragana and katakana are marked with solid lines,
Kanji characters are displayed with double solid lines and underlined.

(5) The operator visually detects recognition errors from the displayed character string and the character type information added thereto.

(6) Display all recognition result candidates for the incorrectly recognized character and let the operator select the correct character. In this case as well, the character type is determined for each candidate character, and the display is performed using a display method according to each character type.

Next, we will discuss a specific example of character recognition in this example in the second section.
This will be explained using figures. Suppose now that a character string written entirely in katakana is recognized, and that the following is obtained as the first candidate for the recognition result.

・The first character ``mouth j'' is a kanji with a very similar shape.
It was mistakenly recognized as "kuchi".

・The second character "-" was misrecognized as "(hyphen/minus)", an alphanumeric character with a very similar shape.

・The third character "ma" was correctly recognized.

At this time, the character recognition device of this embodiment first performs the character recognition process shown in FIG.
Display as follows. That is, a double solid line is added to K for ``口'' to indicate that it is a kanji character. Similarly, “
” has a dotted line indicating that it is an alphanumeric character.
A solid line is added to "ma" to indicate that it is Hiragana/Katakana.

Next, the operator visually checks the displayed character string to try to find any recognition errors. First, if we focus on the first character, it is difficult to determine from the character shape that the displayed character for "mouth" is not the correct answer candidate, "mouth".
In the character recognition device of this embodiment, the character type information given to the character is supposed to be a solid line indicating hiragana/katakana, but it is a double solid line indicating kanji, so it is easy to recognize it incorrectly. It can be determined that Therefore, the operator displays all recognition candidates for the first character,
Select the character for ``mouth'' which is the correct answer candidate.

Even in this case, you only need to focus on the candidates with solid lines indicating hiragana/katakana among the displayed candidates, so you can narrow down the candidates by that much, and you can avoid mistakenly writing similar characters in different character types. Less risky to choose. As a result of the above operations, the displayed character string will be as shown in the same figure (b).

The same goes for the second character, and the fact that the displayed character is an alphanumeric hyphen/minus "" instead of a long sound symbol "" in hiragana/katakana is because the character type information is a dotted line instead of a solid line. It can be easily determined. At this time, the operator can correct the recognition result by performing the same operation as for the first character. As a result of the above operations, the displayed character string becomes as shown in FIG.

Through these series of operations, the correct character string "Rome" can be obtained.

As described above, in this embodiment, recognition errors in characters of different character types with similar shapes can be easily discovered and corrected by using the character type information given to each displayed character as a clue.

In addition, in this embodiment, different types of underlines are added as a method of expressing the difference in the character types of the displayed characters, but this can be done by changing the display color for each character type. The difference can be made clearer. In addition, any method may be used as long as it makes a difference in display, such as changing the font or adding specific markings.

In addition, in this embodiment, the character types to be classified are alphanumeric, hiragana/katakana, and kanji, but these may be further subdivided and integrated.
It is also possible to add other character types.

Effects of the Invention In the character recognition device of the present invention, by changing the display method for each character type, the recognition results are displayed in such a way that the character type to which the character belongs can be easily understood visually. This has a great practical effect because errors in recognition of characters of different character types can be easily discovered.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a character recognition device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram showing a specific example of processing in the character recognition device according to an embodiment of the present invention, and FIG. 3 is a conventional character recognition method. It is a block diagram of a device. 1... Image input section, 2... Character recognition section, 3... Character type determination section, 4... Recognition result display section. Name of agent: Patent attorney Toshio Nakao and 1 other person
1 Figure Figure 2 (α) Rome (b) 4-Ma+・4-Kinaka 1-Kohakuhei (Q Koni! Figure 3)

Claims (1)

[Claims] an image input unit that inputs a document as image information; a character recognition unit that recognizes character strings included in the document image input from the image input unit and converts them into character code strings; a character type determination unit that determines which character type the character code belongs to among a plurality of character types classified in advance; and a character type determination unit that determines which character type the character code belongs to among a plurality of character types classified in advance; A character recognition device comprising: a recognition result display unit that displays a character type to which a character belongs so that it can be visually and easily understood by changing a display method.