JPS6240579A - Translating device - Google Patents

Abstract

PURPOSE:To input a sentence at a high speed, and to correct a wrong input by correcting a character which has been inputted by an optical character reader by a character which has been inputted by a keyboard, and also correcting and editing a sentence which has been translated by a translating means. CONSTITUTION:A character on paper is scanned and read by an optical system 1, and a character pattern by a detecting function 2 is detected and also specified by a pattern in a dictionary 3. As for a character which is not specified, a rejection mark is put, and its character is registered i a sentence memory 5 together with a recognized memory. Subsequently, in a word processor mode, the character to which the rejection mark has been put is inputted from a keyboard 9. When a sentence input is completed, in case of there is other character than that which is contained in a dictionary of a translating function 8, among the characters which have been inputted by a translation mode, an alternate character is inputted from the keyboard 9 and registered in the sentence memory 5. When all the inputted characters correspond to characters of the dictionary of the translating function 8, the translation processing is executed. A correction and editing of a translated sentence are executed by the keyboard 9 by selecting the word processor mode.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a translation device, and particularly to a translation device that can easily and appropriately input a sentence to be translated, and can correct an output translated sentence. Regarding equipment.

(Summary of the Invention) The present invention provides an optical character reading device I (
OCR, 0ptical Character Rea
der) and the correction/editing function of the word processor to quickly input sentences, correct incorrect inputs, and correct/edit inappropriate output translated sentences.

(Prior Art) Machine translation systems that use a knowledge base and language-independent intermediate representations, and multilingual machine translation systems that use language-independent conceptual structures as the basis of intermediate representations, are generally known. There is. However, in this type of translation system,
The text to be translated was entered using a keyboard.

(Problems to be Solved by the Invention) Therefore, there are problems in that text input using the conventional translation device is slow due to keyboard input, and correction in case of incorrect input is troublesome.

Furthermore, there is a problem in that if an inappropriate translated sentence is output due to an input error or a defect in the translation system, it cannot be corrected.

In view of the above-mentioned problems, the present invention provides a translation Rfft system that allows for quick character input (and easy correction of erroneous input and inappropriate output translated sentences).
The purpose is to provide

(Means for Solving the Problems) In order to solve the above-mentioned problems, the present invention provides a means for optically reading characters, a means for recognizing characters input by the reading means, and a method for inputting characters. a keyboard for correcting the characters input by the reading means with the characters input by the keyboard;
! 14'e! Having a means for correcting and editing the text translated by the translation means! Let it be i.

(Example) Hereinafter, the present invention will be described in detail with reference to the drawings. FIG. 1 is a small-sized schematic configuration diagram of an embodiment of a translation device according to the present invention.
2 is a schematic perspective view of the apparatus of FIG. 1; FIG.

This device includes an optical system 1 composed of COD, etc., and an OCR
It has a detection function 2 and an OCR dictionary 3. The optical system 1 can read and input characters already printed or written on paper or the like, and may be of a relatively inexpensive hand scan type.

The OCR detection function 2 corrects the data from the optical system 1, performs pre-processing such as normalization and thinning, and then calculates the slope, length, etc. of the line.
Features such as region shape can be extracted, but the ability to do so may be relatively small.

The OCR dictionary 3 includes basic pattern information such as straight lines and curved lines of characters (4T), partial pattern information such as single character Ben and Tsukusu, alphanumeric characters, hiragana, and katakana.

Character pattern information such as kanji characters is stored, but the storage capacity may be relatively small.

The OCR detection described above uses the matching function 4, which matches the pattern detected by function 2 with the pattern in the dictionary 3, and stores the matched characters, etc. A document memory 5 for storing documents, a display section (CRT) 6 (Fig. 2), and a correction/editing function (for performing kanji conversion, document creation, proofreading, addition, correction, insertion, etc.)
A word processor) 7, a translation function 8 having a translation dictionary, and a keyboard 9 (FIG. 2) for selecting OCR mode, corrective editing (word processing) mode, or translation mode, and for inputting characters. 10 is interf]-
-S (1/F).

Next, the operation of the translation device having the above configuration will be explained with reference to FIG. Here, operator tit optical system 1
Alternatively, characters can be entered arbitrarily using the keyboard 9, but in order to explain the features of the present invention more clearly, characters printed or scratched on paper etc. are first input in OCR mode, and then characters are read in OCR mode. This section explains how to input characters that cannot be entered using the corrective editing function (Warb[] mode).

In this case, the operator first selects the OC using the keyboard 9.
The R mode is selected and the characters on the paper are scanned and read by the optical system 1.

First, in Sudetsubu (hereinafter referred to as ST) 1, the OC
It is determined whether or not the R mode is selected, and since the OCR mode is selected as described above, the process proceeds to ST2.

In ST2, the OCR detection function 2 digitally converts the data of one character read by the optical system 1.

M sound removal, distortion correction, blur correction 1 image enhancement, normalization,
After pre-processing such as cropping, binarization, and thinning, line extraction, region extraction, line slope/length, etc.

Character patterns are detected by extracting features such as region shapes.

In ST3, the detected character pattern is compared with the basic pattern of the OCR Dictionary 3, and with partial patterns such as oddities and character patterns.

In ST4, it is determined whether the detected character pattern can be specified by the pattern in the dictionary 3. If the character pattern cannot be specified, the process proceeds to ST5, and if it can be specified, the process proceeds to 8-cho6. In ST5, the characters read are judged to be mutually exclusive.
t) The mark is recorded in the document memory 5, and the recognized characters are registered in the document memory 5 in ST6.

Thereafter, when the operator scans each character on the paper with the optical system 1, the apparatus repeats ST2 to ST6 to perform pattern extraction, collation, and memory zero recording. In this case, it is thought that many of the misread or rejected characters are complex characters such as kanji.

Next, when the operator selects warp [commode] using the keyboard 9, the apparatus advances from ST1 to ST7. At this time, the OCR mode (ST1 to ST6) is displayed on the CRT6.
The sentences entered in and registered in the sentence memory 5 are displayed.

Continuation < a-3 in ST8, rejection (Rej
eCt) Determine whether or not the turn is registered;
Reject ST9 if registered
The process proceeds to the processing routine (FIG. 4); on the other hand, if it is not registered, the process proceeds to 5TIO.

To explain the rejection processing routine in FIG.
At T21, of the text displayed on the CRT 6, the area of the rejected characters is displayed in a flickering manner using a block cursor. At the subsequent step 5T22, an instruction to insert the rejected characters is displayed on the CRT 6.

When the operator inputs characters using 4--board 9,
This device registers this character in the corresponding area of the rejection mark in the text memory 5 and returns to ST8. In 5T23, further rejected in sentence memory 5? - It is determined whether or not the device is registered, and if it is registered, the process returns to 5T21, and if it is not registered, the process returns to the main flow of 5TiO. That is, even if an inexpensive OCR with small functions is used, it is possible to easily input simple characters.

In the 8th block 10, the text correction/editing process designated by the ■-board 9 is performed. Here, the operator selects OCR mode (S
It is possible to edit the characters input in TI~6) and to correct misread characters.

When the text input is completed as described above and the operator selects the i-translation mode on the 1-board 9, the apparatus proceeds from ST1 to 5T7-) ST11→5T12. At 5T12, if a rejection mark is still registered in the text memory 5, the process advances to 5T13 and the same processing is performed according to the routine shown in FIG. Incidentally, here, the text is continuously displayed on the CRT 6.

Continuation<In ST14, among the characters input as described above,
It is determined whether there are any characters other than those in the dictionary of the translation function 8, that is, whether there are any corresponding translated characters. If the corresponding translated character is not in the dictionary, proceed to 5T15 and 8
The characters entered in 1 to 10 are displayed on the CRT6 with a block cursor inverted and flickered, and then 5T1 is displayed.
At step 6, an instruction to insert an alternative character is displayed. When the operator inputs a substitute character using the 1-board 9, this device converts this character into Kanji and registers it in the corresponding 1-7 of the document memory 5. That is, in 5T14-16, characters inappropriate for translation can be corrected. If all the input characters correspond to the characters in the dictionary of the translation function 8, the process advances from 5T14 to 5T17 and translation processing is performed.

There is a keyboard 9 on the platform where the operator checks the translated text.
Select word processing mode. Then, this device
TI to ST 7-8 T 8→ST9→ST'10
The translated text is displayed on the CRT6.

The operator uses the keyboard 9 to edit the translation text while observing this display. Therefore, the present invention can correct erroneous inputs and inappropriate inputs, as well as correct and edit output translated sentences.

In addition, the present invention includes online handwritten character reading (0n-1, i
It is also possible to add other functions (ST18) such as neCharacter Reading).

(Effects of the Invention) As explained above, the present invention has an optical character reading function and a text correction/editing process (WordBrowsetsuIt'), which increases the text input speed, corrects incorrect input, and corrects inappropriate input. It becomes possible to correct and edit the output translated text.

[Brief explanation of drawings]

FIG. 1 is a schematic configuration diagram showing one embodiment of the translation device according to the present invention, FIG. 2 is a schematic perspective view of the device shown in FIG. 1, and FIG. 3 is a diagram for explaining the operation of the device shown in FIG. 1. Flowchart FIG. 4 is a flowchart of the rejection processing routine of FIG. 3. 1...Optical system, 2...OCR detection function, 3...O
CR dictionary, 4... Verification function, 5... Document memory,
6...CRT, 7...Correction editing function, 8...Translation function, 9...Keyboard. Patent applicant: Tateishi Denshu Co., Ltd. Agent Patent attorney: Tetsuji Iwakura (1 other person) Figure 1 Figure 2

Claims (1)

[Claims] means for optically reading characters; means for recognizing characters input by the reading means; a keyboard for inputting characters; a translation means; A translation device comprising means for correcting the text translated by the translation means as well as for correcting and editing the text translated by the translation means.