KR20080041974A - Method and apparatus for entering text in a mobile device - Google Patents

Abstract

A method and an apparatus for inputting characters at a mobile terminal is provided to enable a user to input characters without overlapping by using a character inputting scheme based on a language modeling algorithm. A method for inputting characters at a mobile terminal includes the following several steps. A processor checks whether a mobile terminal is in a word inputting mode(101), and the processor checks spellings in correspondence with keys pushed by a user(102). The processor searches for all the possible words which can be made from the spellings inputted to that time and checks whether there exist available words by using the language modeling algorithm(103). The processor checks whether at least one possible word is found on the basis of the algorithm(104). If so, the processor continues to calculate probability on each word which can be made from text inputted by the user(106). The processor presents a word or an expression with the highest probability to the user(107). The processor checks whether the presented word or expression is accepted by the user(108). The processor checks whether the user selects a wanted word in a presented list(109). The processor selects a similar word and completes a wanted word by enabling a user to add other characters to the selected similar word(111).

Description

Method and device for entering text in a mobile terminal {Method and Apparatus for entering text in a mobile device}

1A is a schematic front view of a mobile terminal according to an embodiment of the present invention.

FIG. 1B is a schematic structural diagram of some elements of the mobile terminal of FIG. 1A; FIG.

1C is a schematic diagram of a numeric keypad of the mobile terminal of FIG. 1A;

2 is a flowchart illustrating a character input method according to an embodiment of the present invention.

3 is a flowchart showing an example of a character input method according to another embodiment of the present invention;

The present invention relates to a character input method and apparatus, and more particularly, to a method and apparatus for inputting a character in a mobile terminal.

A method of communicating by using text input from a keypad of a telephone is widely used. Among the services providing communication using the above-mentioned text, a global system for mobile communication (hereinafter referred to as 'GSM') and a universal mobile telecommunications system (hereinafter referred to as 'UMTS') Text messaging services such as the Short Message Service (hereinafter referred to as SMS) provided by the network are preferred.

The most common way of entering text is the so-called multi-press method (also called 'tap' method), Tegic Communications 'T9' character input ('T9'), Motorola There are predictive character input schemes such as the iTAP ^™ intelligent keypad input system, and Zi's eZiText ^™ .

In a typical telephone keypad, groups of letters in alphabetical order are associated with numeric keys. This association is shown in FIG. 1C. For example, the letters 'a', 'b' and 'c' are associated with the number '2'. Thus, a single press of the numeric key has a multiplicity, since one numeric key can represent one of the sets associated with three or four spellings.

The duplicate-pressing method is a method of tapping the same key a plurality of times, thereby eliminating the problem of the above-mentioned meaning. The user presses the key corresponding to the position of the characters arranged in the order of Standard a plurality of times. For example, for the number 2 key, the user presses 'a' once, and presses 'b' twice.

The disadvantage of a redundant-pressing system is that more than one key press is required per spelling.

Accordingly, the present invention is directed to ameliorating the above-mentioned disadvantages.

The present invention also provides an improved method and apparatus for entering text in a mobile terminal.

The present invention also provides a method and apparatus for inputting characters using language modeling.

According to one aspect of the present invention, a method for inputting a character in a mobile terminal is a process of receiving an input of a key representing one or more alphabetic, alphanumeric spellings, and spelling corresponding to the input keys. Searching for possible words or expressions to include, and determining a word or expression having the highest probability among the searched words or expressions using language modeling.

According to another aspect of the preferred embodiment of the present invention, the method for inputting a character in the mobile terminal is a process of receiving a key (Key) indicating one or more alphabetic, alphanumeric spelling and spelling corresponding to the input key; Searching for possible words or expressions to include, and determining the most likely corresponding word or expression in consideration of the context and / or training text of the input text.

According to another aspect of the preferred embodiment of the present invention, a device for inputting a character in a mobile terminal includes a keypad having keys for representing one or more Hangul, alphabet and numeric spellings, a memory for storing words or expressions; Search for possible words or expressions that include spellings corresponding to keys entered through the keypad, and determine the highest probability word or expression among the searched words or expressions using language modeling; It includes a processor.

According to another aspect of the preferred embodiment of the present invention, a device for inputting a character in a mobile terminal includes a keypad having keys representing one or more Hangul, alphabet, and numeric spellings, a memory for storing words or expressions; Search for possible words or expressions that include spelling corresponding to keys entered through the keypad, and take into consideration the context and / or training text of the text that includes the entered words or expressions, And a processor for determining a word or expression having a correspondence of probability.

The foregoing has outlined rather broadly the features and technical advantages of the present invention in order that those skilled in the art may better understand it from the following detailed description of the invention. In addition to these features and advantages, further features and advantages of the present invention which form the subject of the claims of the present invention will be better understood from the following detailed description of the invention.

DETAILED DESCRIPTION A detailed description of preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that the disclosed concept and specific embodiments of the invention described below may be changed or modified to achieve the technical problem to be achieved by the present invention. Those skilled in the art will also recognize that concepts and structures equivalent to the concepts and structures disclosed herein do not depart from the spirit and scope of the broadest form of the invention. In the following description of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

The present invention proposes a predictive character input method using a probability calculation method as an improvement method of the conventional overlap-pressing system.

The predictive character input method is a method in which a user can input a character by pressing only one key per spelling. When a user types a word one letter at a time, the system automatically compares all possible combinations that can be derived from a series of input keys with lexical words and 'infers' the spelling words that you want to type. do. However, many dictionary words share the same numeric codes, so in this case the system presents alternative spellings and words in the list to the user. Then, the user selects a word to be input by the user from the list.

Systems that support predictive character input methods select and display the most prominent words most frequently used in a specific language, and list other words that share the same numeric code (in order of least likely occurrence). . This also reduces the need to interrupt further character input to select one of the words in the list if the word selected is the most likely solution as the most likely solution.

The use of predictive character entry systems can make character entry much more efficient.

Users often want to enter non-dictionary words, such as names, slang or abbreviations, especially when entering SMS characters. In this case, non-dictionary words may be input using a predictive character input system, for example, by using previously used text or directly input by the user.

1A is a schematic diagram of a mobile terminal 10. The terminal 10 includes a display unit 26, a microphone 16, speakers 18, a keypad 21, an antenna device 20, and navigation keys. keys) 23. Herein, the mobile terminal 10 can be easily carried by an individual such as a mobile phone, a personal digital assistant (PDA), a portable multimedia player (PMP), and the like, and can perform functions such as information processing, information provision, and communication. Branches cover electronics.

1B, the terminal 10 includes a processor 22, a wireless device 24 for communicating with other devices via a mobile communication network, an antenna device 26, a memory 28, and a user interface ( User Interface (UI) device 30 is configured. A Subscriber Identity Module (SIM) 32 for GSM terminals or a Universal SIM (USIM) 32 for UMTS mobile terminals is typically inserted into the mobile terminal to provide services over a mobile communication network. Can provide. The SIM or USIM includes a memory clip or a microprocessor.

FIG. 1C schematically illustrates the numeric keypad of the mobile terminal of FIG. 1A.

Each numeric key 40 is mapped to a respective number 42 followed by one or more associated spells 44. For example, 'A', 'B' and 'C' are usually mapped to number key '2' with the number '2', and 'D', 'E' and 'F' are the number '3' Mapped to number key '3'.

In this way, alphanumeric characters Text can be input using the numeric keypad 21. Although only alphabetic letters and numbers are shown in FIG. 1B, letters of other languages including Korean can also be mapped to respective numeric keys 40 by hardware or software assignment. For example, in the case of Hangul, 'ㅣ' can be mapped to the number '1' key, and 'ㄱ', 'ㅋ' can be mapped to the number '4' key according to Cheonjiin's method of Hangul arrangement. .

As a character input system, software applications that typically drive the processor 22 of the terminal are used. The prediction systems have a dictionary for automatically comparing all possible combinations that can be derived from input keys with dictionary words, and for inferring spellings and words to enter, such a dictionary having a memory (28). Usually stored in).

First embodiment

In order to provide an improved character entry system for a keypad with a small number of keys, for example a mobile terminal with twelve keys, one letter to be selected from among a number of letters associated with one key is selected. An improved method is proposed.

Hereinafter, embodiments based on language modeling will be described in detail. Here, language modeling is an algorithm that infers a word from spelled input by calculating the probability of a trained word or a string of words from a dictionary.

A system supporting character input is provided as a software application stored in a memory of a mobile terminal and driven by a processor.

A description with reference to FIG. 2 is as follows. In operation 101, the processor 22 of FIG. 1B checks whether the user inputs a character into a mobile terminal. In step 102, the processor 22 checks spelling corresponding to one or more keys pressed by the user. Here, the spelling may be inferred by confirming a key pressed by the Hangul, alphabet, or numeric spelling that the user wishes to input. In step 103, the processor 22 searches for all possible words that can be written in spelling corresponding to the keys pressed by the user by using a language modeling algorithm and checks whether there are possible words.

In step 104, the processor 22 checks whether at least one or more possible words are searched using the algorithm. If only one word is found as the corresponding word, the processor 22 displays the one word on the display unit in step 105. In step 116, the processor 22 checks whether there is an input representing that the displayed word is a word to be input by the user. Here, the expression that the displayed word is a word that the user wants to input may be expressed by the user pressing a non-alphabetic numeric key. The non-alphabetic numeric key refers to specific keys such as '1' key, '*' key, '0' key, and '#' key of FIG. 1B, and the word input completion or entry into the spell input mode. It is used to request. If the displayed word is input by the user, the processor 22 approves the word in step 117, and then returns to step 101 to allow the new word to be input continuously. If the word displayed in step 116 is not a word desired by the user, the processor 22 proceeds to step 115 and enters a spell input mode to be input by the user.

On the other hand, if one or more corresponding words are retrieved in step 104, the process 22 continues to calculate a probability for each word that may be made from the text entered by the user in step 106.

These probabilities are computed using a Prediction by Partial Matching (PPM) algorithm. The PPM counts the number of times words and sequences of words occur in 'training text'. Such 'training text' is composed of a combination of 'previous training text', 'previous use text' and 'idiom text' which will be described later.

'Previous Training Text' is text stored on the device before the algorithm is used by the user.

'Previous user text' is text selected by the user to train the character input system.

'Organic text' is text registered as user input.

Any of the three training texts described above for combinations of texts may be used. The 'idiom text' is particularly useful because character input can be dynamically adapted to the user to provide a self-improvement character input system.

Training texts used in the algorithm are stored in the memory of the mobile terminal.

By the algorithm in step 107, the processor 22 suggests the most likely word or expression to the user, and presents it to the display by displaying a list of one or more words found in step 104. The words are suggested in order of decreasing probability of occurrence, starting with the most likely words following the proposed most potent word or expression.

In step 108, the processor 22 checks whether there is an input representing that the word which is likely to be input by the user. In this case, the expression that the probable word is a word that the user wants to input may be expressed by the user pressing a non-alphabetic numeric key. If the probable word is an input of a word to be input by the user, the processor 22 proceeds to step 117 to approve the word.

If the user's desired word is a word in the list proposed in step 107, the user can select from the list of the list using the navigation keys of the keypad, or press the associated key to select one of the words displayed in the list. By pressing you can enter one or more other Hangul, alphabet, or numeric spellings. Accordingly, process 22 verifies in step 109 that the word desired by the user is selected from the list by the user. If a word is selected, process 22 proceeds to step 117 to accept the selected word.

Also, it is checked whether a word similar to a word desired by the user exists in the list proposed in step 107. If present, processor 22 selects the similar word in step 111 and then completes the desired word by entering one or more other Hangul, alphabet, or numeric spellings. Here, the similar word refers to a word that is similar to a word to be input by the user and can be completed by correcting one or more spellings.

However, if the word desired by the user does not exist in the list, the processor 22 enters the spelling input mode to be input by the user in step 115, which will be described later in more detail.

If no word corresponding to the keys entered so far is found in step 103, the processor 22 searches for possible words starting with words that can be written with a given set of keys using an algorithm in step 112. If a possible word is found, in step 113, the processor 22 checks whether one or more words are found.

If only one word is retrieved in step 113, the processor 22 proceeds to step 105 to display the single word on the display.

If one or more words are retrieved in step 113, the processor 22 calculates the probability of occurrence for each word using an algorithm in step 114. Here, the probability is calculated by suggesting a complete word of the highest probability or a sequence of spells with the largest sum of the probabilities of the words that complete the sequence of spelling.

In step 107, the processor 22 suggests a word or sequence consisting of spellings having the highest probability displayed on the display unit along with a list of other corresponding words as described above.

If a possible word is not found in step 112, the processor 22 proceeds to step 115 and enters a spelling input mode to be input by a user.

If the user inputs a non-alphabetic numeric spelling at any time, the processor 22 interprets in step 101 that each key associated with the entry of a new word has been entered.

The spelling input mode starts when the user presses a set of keys, such as in step 110, 112, or 116, that the word of the first entered spelling that can be given by this set of keys is not searched, or suggested It starts when all of the words do not correspond to the word you want to enter.

Other versions of spell input modes may be applied, including spell input modes in the following examples.

Spell entry mode starts with a blank screen and allows the user to enter a word by means of a duplicate-pressing character entry system. Alternatively, spelling input mode uses a set of keys entered up to that time to suggest a set of spellings that the user can edit or complete.

If words entered in the spelling input mode are added to a dictionary (eg, 'training text') and are often used, they may have a higher probability than words that existed previously.

Several examples are given below to illustrate the foregoing.

Example 1

Start from a blank screen

User presses key two: 'ABC'

'A' is a single word, 'B' is not a single word, and 'C' is not a single word.

=> Suggest a on screen

User presses key six: 'MNO'

Possible words are 'AN', 'AM', 'CO', 'CM', and 'AN' is used most often.

=> Suggest AN on the screen

User presses key three: 'DEF'

Possible words are 'AND', 'COD', '2ND', and 'AND' is used most often.

=> Suggest AND on screen

Example 2

Start with a screen with 'A' in the empty space

User presses key two: 'ABC'

'A' is a single word, 'B' is not a single word, and 'C' is not a single word.

=> Suggest a on screen

User presses key six: 'MNO'

Possible words are 'AN', 'AM', 'CO' and 'CM'. This time the word 'AN' is not suitable after 'A', and 'CM' is most suitable after 'A'.

=> Suggest 'CM' on screen

User presses key three: 'DEF'

Possible words are 'AND', 'COD' and '2ND'. Again, the word 'AND' is not suitable after 'A', and '2ND' is most suitable after 'A'.

=> Suggest 2ND to the screen (T9 suggests AND)

Example 3

Start with a screen that says 'I DON'T KNOW'

User presses 4 and 3 keys

Here T9 would suggest 'HE', but the system of the present invention generally learns that 'IF' is most suitable after 'KNOW' even though the word 'HE' is more commonly used than the word 'IF'. come.

=> Suggest IF on screen

Example 4

Start with a screen that says 'GONE'

User presses 4, 6, 6, and 3 keys

Here T9 would suggest 'GOOD', but the system of the present invention generally learns that 'HOME' is most suitable after 'GONE' even though the word 'GOOD' is more commonly used than the word 'HOME'. come.

=> Suggest HOME on screen

In the first embodiment, the sequence of key presses and context, e.g., taking into account previous words in the text being written, determine how the probability for each word can be written in a sequence of spelling. Explained. In conclusion, the word with the highest probability is suggested to the user. The probability for each word is computed using PPM and depends on the number of occurrences of words in one or more training texts and the context of those occurrences (eg, previous words).

Second embodiment

The second embodiment, which is an extended version of the first embodiment, will now be described. In addition to the features described above, systems supporting character input include the additional feature that the probability of the word being written is dependent not only on the context (eg, previous words) but also on the current situation.

The system takes into account situation dependent parameters to determine the probability for each word that can be written with a given sequence of keys.

One way to take the context dependent parameters into account is to use other training texts that are context dependent.

For example, the system may use the specific training text for each recipient of the text being input or the specific training text for each group of recipients, such as friends, family members, business associates, and the like.

Alternatively, other training texts may be used depending on the settings or profile currently being used at the mobile terminal. For example, the first training text can be used as a profile used when in an office or during a meeting, and the other training text can be used as a profile used during free time.

Another alternative is to use different training texts depending on the time of day and / or day of the week.

For example, the first training text is used between 9 am and 6 pm on weekdays, and the second training text is used on weekends or after 6 pm on weekdays.

Note that the situation dependent use described above can be combined in any other possible way.

Instead of using different training texts for different situations, the training texts can be shared between situations by assigning different weights to the respective training texts. For example, higher weights are given for the occurrences of words in the training text that are specific to the current situation, and are calculated higher, whereas normal occurrences are calculated for the occurrences of words in other training texts. In this way, frequent situations have large training texts, and the probability computed for those situations is correct. However, situations with less training texts may use information from other similar situations until less training texts are large, since the probability for those situations cannot generally be calculated accurately.

Hereinafter, a flowchart of an example will be described with reference to FIG. 3.

In this example, the user is assumed to be a merchant. The processor 22 of FIG. 1B confirms that the number keys 4, 6, 5, and 3 are pressed by the user through the keypad of the mobile terminal in step 201. In step 202, the processor 22 determines that words that can be written by a given sequence of keys are 'hold', 'gold', 'hole' and 'golf'. If a message containing the sequence is written on a Monday morning to be sent to a company colleague (step 203), the processor 22 uses the first training text to determine the probability (step 204). In step 205, the processor 22 determines that the word 'gold' has the highest occurrence probability among the possible words determined in step 202. Thus the word is suggested to the user.

On the other hand, if a message to be sent to a friend and Golf Partner is written on a Friday evening, the processor 22 uses the second training text (step 207). Using the second training text, processor 22 determines that the word 'golf' is most likely to occur (step 208) and suggests the word to the user when the user enters the text.

In the embodiment described above, the situation for computing the situation dependent probabilities includes the recipient or group of recipients, profile or settings of the message being created, or the time of day or day of the week. However, note that other situations may also be considered when calculating situation dependent probabilities, such as, for example, situations derived from calendar entries.

Although the PPM is described as an example of a technique for determining probabilities in the above-described embodiments, of course, other techniques may also be used.

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications are possible without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the scope of the following claims, but also by those equivalent to the scope of the claims.

As described above, the present invention has an advantage in that a character input method using a language modeling algorithm in a mobile terminal can input characters without pressing keys repeatedly.

Claims (42)

In the method for inputting characters in the mobile terminal, Accepting input of keys representing one or more alphabetic, alphanumeric spellings, Searching for possible words or expressions that include spelling corresponding to the input keys; And determining a word or expression having the highest probability among the searched words or expressions using language modeling. The method of claim 1, wherein the word or expression determined in the determining process is displayed to the user. The method of claim 2, wherein the probability is a probability of generating the words or expressions and is a criterion for determining the words or expressions searched in the searching process. The method of claim 3, wherein the words or expressions found in the searching are displayed to the user. The method of claim 4, wherein the words or expressions searched in the searching process are arranged in order of occurrence probability. The method of claim 5, wherein a word or expression to be input by the user is selected from among the displayed words or expressions. The method of claim 6, wherein one of the possible words or expressions is selected by the user. 7. The method of claim 6, wherein after the selection of one of the words or expressions, a subsequent Hangul, alphabet, and numeric spelling is further input. The method of claim 8, wherein words or expressions are determined by a word or expression corresponding to an input key sequence. The method of claim 1, wherein the words or expressions determined in the searching process are determined to start with a sequence of spellings corresponding to the input key sequence. The method of claim 10, wherein the word or expression having the highest probability in the determining process is a complete word or expression having the highest occurrence probability. The method of claim 10, wherein the word or expression having the highest probability in the determining process is determined as a sequence of spells that complete the word having the largest sum of the probabilities of the words among words completed using the input sequence. Character input method characterized by the above-mentioned. The method of claim 12, wherein the probability of the word in the determining process is calculated by a prediction of partial matching algorithm. The method of claim 13, wherein the occurrence probability of the specific word or expression is determined from the training text. The method of claim 14, wherein the training text is at least one of text stored in a mobile communication device, text selected by a user for training purposes, and text input by a user before inputting spellings into the mobile terminal. Character input method. 16. The method of claim 15, wherein training text is used to determine a probability for the word or expression in the determining. 17. The method of claim 16, wherein additional parameters are considered when determining a word or expression with the highest probability in the determining process. 18. The method of claim 17, wherein the additional parameters are context dependent parameters. 19. The method of claim 18, wherein the additional parameters include any one or more of a recipient, a group of recipients, selected settings or profile of a mobile terminal, a text entry date, a text entry time, and a text entry day. . 20. A method according to claim 19, wherein the additional parameters are taken into account using different training texts. 21. The method of claim 20, wherein in the determining process, the probabilities are determined using the training texts, and different weights are assigned to the training texts in consideration of the additional parameters. In the method for inputting characters in the mobile terminal, Receiving input of a key representing one or more alphabetic, alphanumeric spellings, Searching for possible words or expressions that include spelling corresponding to the input keys; And determining a word or expression having a highest probability correspondence in consideration of a context and / or training text of the text including the input words or expressions. In the device for inputting characters in the mobile terminal, A keypad with keys representing one or more alphabetic, alphanumeric spellings, Memory for storing words or expressions, A processor that searches for possible words or expressions that include spelling corresponding to keys entered through the keypad, and uses language modeling to determine the highest probability word or expression among the retrieved words or expressions Character input device comprising a. The character input apparatus of claim 23, wherein the processor displays the searched or determined word or expression to a user. The character input device of claim 24, wherein the probability is a probability of generating the words or expressions and is a criterion for determining the searched words or expressions. 27. The character input device of claim 25, wherein the searched words or expressions are arranged in order of occurrence probability. 27. The character input device of claim 26, wherein any one of the possible words or expressions is selected by the user. 28. The character input device of claim 27, wherein the processor is further configured to further input Korean, alphabet, and numeric spelling after one of the words or expressions is selected. 29. The apparatus of claim 28, wherein the words or expressions are determined by a word or expression corresponding to an input key sequence. 29. A character input device according to any one of claims 23 to 28, wherein said determined words or expressions are determined to begin with a sequence of spellings corresponding to an input key sequence. 31. The character input device of claim 30, wherein the highest probability word or expression is a complete word or expression with the highest probability of occurrence. 32. The apparatus of claim 31, wherein the word or expression having the highest probability is determined as the word or expression having the largest sum of probabilities among words or expressions including the input sequence. 33. The apparatus of claim 32, wherein the probabilities for the words are calculated by a prediction of partial matching algorithm. 34. The character input device of claim 33, wherein a probability of occurrence of a particular word or expression is determined from training text. 35. The method of claim 34, wherein the training text is at least one of text stored in a mobile communication device, text selected by a user for training purposes, and text input by a user before inputting spellings into the mobile terminal. Character input device. 36. The character input device of claim 35, wherein the probability for the word or expression is determined using training text. 37. The character input apparatus of claim 36, wherein the highest probability word or expression is defective in view of additional parameters. 38. The character input device of claim 37, wherein the additional parameters are context dependent parameters. 39. The text input of claim 38, wherein the additional parameters include any one or more of a recipient, a group of recipients, selected settings or profile of a mobile terminal, a text input date, a text input time, a text input day, and the like. Device. 40. The character input device of claim 39 wherein the additional parameters are considered using different training texts. 41. The character input device of claim 40 wherein the probabilities are determined using the training texts and different weights are assigned to the training texts to account for the additional parameters. In the device for inputting characters in the mobile terminal, A keypad with keys representing one or more alphabetic, alphanumeric spellings, Memory for storing words or expressions, Search for possible words or expressions that include spellings corresponding to keys entered through the keypad and have a highest probability in view of the context and / or training text of the text containing the entered words or expressions And a processor for determining a word or expression having a corresponding relationship with the word.

Images