WO2007048213A1

WO2007048213A1 - Method of formation of a word by inputting characters by means of a numerical keypad

Info

Publication number: WO2007048213A1
Application number: PCT/BR2006/000229
Authority: WO
Inventors: Silvio Mano Maeta; Paulo Sirum Ng; Ivandro Sanches
Original assignee: Genius - Instituto De Tecnologia
Priority date: 2005-10-25
Filing date: 2006-10-25
Publication date: 2007-05-03
Also published as: BRPI0506037A; WO2007048213A8

Abstract

The present invention generally refers to a method of inputting words in a screen from a numerical keypad. According to an embodiment of the invention, the method comprises the steps of (i) from a training lexicon, to define the contextual questions used to map the characters corresponding to the keys pushed in a numerical keypad; (ii) to register each push of the key in the keypad that corresponds to the group of characters that contains the character that is desired to insert in a screen; (iii) to analyze, at each push of the key, the context of the keypad keys pushed before and after each one of the pushed keys; (iv) to show, at each push of keypad key, the corresponding character to each pushed keypad key on the basis of the contextual questions stored in memory, said characters being shown in a screen; and (v) to finish the formation of the word when any corresponding keypad key corresponding to the punctuation, space or any special character is pushed.

Description

"METHOD OF FORMATION OF A WORD BY INPUTTING CHARACTERS BY MEANS OF A NUMERICAL KEYPAD"

FIELD OF THE INVENTION

The present invention generally refers to a method of inputting words in a screen of an electronic device from a numerical keypad.

More specifically, the present invention refers to a method of inputting words from a numerical keypad that permits to reduce the number of times that the keypad is pushed to insert the desired word, in addition to reduce the necessary space of memory of the electronic device for the insertion of said words.

DEFINITIONS

For the present specification, the following definitions or abbreviations are used:

Character: a graphical symbol that, alone or in association with other symbols, is capable of establishing a meaning to a human being, even if said meaning is associated with the regional culture, language, or customs, such as words, abbreviations, punctuations, symbols etc.

Word: the association of one or more characters that defines a semantic meaning in a specific language. Numerical keypad or keyboard: a keypad comprising a numerical sequence, such as a telephone keypad, and a series of characters associated to each one of the ten numbers of the decimal base (0 the 9).

To push or to press, and its conjugations: refers to the action of pushing and releasing a key of a numerical keypad. Lexicon: compilation of vocables of a language, related to a definite criteria.

GSM: Global System for Mobile Communications. SMS: Short Message Service. KSPC: Keystrokes Per Character.

BACKGROUND OF THE INVENTION

Presently, the operators of mobile or cellular telephony provide a service known as Short Message Service or SMS that is based on the functionality of sending and receiving text messages by means of a cellular phone. The SMS was originally developed as part of phase 1 of the GSM standard.

The text of the message may contain words, numbers, symbols, or combinations of these elements, and the input of the characters of the message is carried out through the numerical keypad of the cellular phone. It is well known that the keypad of cellular phones contains letters associated to the numbers of each key. For example, the keypad key of the number "2" contains the letters "a", "b" and "c". In the same way, the keypad key of the number "3" comprises the letters the "d", "e" and "f , and thus successively until the key of the number "9". The other keys, such as the keypad key of the number "1", the keypad key of the number "0", the asterisk (*) and the pound (#) generally are destined to the insertion of other types of characters in the text of the message, such as space, diverse definition of capital or small letters, punctuation, commas or even other symbols. When the user wants to insert a word in the text of a message, it is necessary to push an adequate number of times the numerical key, in the same order the letter appears in the key, in order to input the desired letter to form the word. For example, to insert the word "home" the user needs to push the keypad key of the number "4" two times for the letter "h", three of the number "6" to insert the letter "o", one more time for the letter "m", and finally two more times the key of the number "3" for the letter "e", thus forming the word.

Beyond demanding a series of push for inputting the letters in order to form the word, it is also necessary to wait a moment when it is necessary to insert more than one letter associated to the same keypad key. For example, if the user desires to insert the word "no", it is necessary to push two times the keypad key of the number "6" for inputting the letter "n", wait a moment, and to push again three times the same key for inputting the letter "o". The technique of inputting words according to the method defined above is known as Multitap, and it is described, for example, in the patent US 6,204,848. The disadvantage of this technique is the need of a great number of push of the keypad keys to form a word, beyond the necessity of waiting a moment when it is desired to insert more than one letter associated to the same numerical key.

A measure to define the average number of push in a numerical keypad key to insert a letter or character is called Keystrokes Per Character (KSPC). For example, the method Multitap described above has a KSPC of about 2.03. Aiming to reduce the number of push of a numerical keypad key for the insertion of characters, and consequently to reduce the KSPC index, other techniques for the insertion of characters by means of a numerical keypad had been developed, such as the T9, of Tegic Communications, iTAP, from Motorola, and eZiText, from ZiCorporation. The T9 technique is described, for example, in US 6,307,549 wherein a dictionary is stored in the memory and a table of frequency of use is employed to predict the typed words.

The patent US 5,952,942 describes, for example, the technique iTAP, wherein a recorded dictionary in the memory is composed of two parts: a fixed one, implanted in the memory, and a variable part that, in accordance with the frequency of use of one determined word by the user, adds or removes words.

The ZiText technique is described, for example, in patent US 6,801 ,659, wherein beyond using the technique Multitap, it is also used a prediction associated with a dictionary, as much for ideographic symbols as for alphabetical symbols.

However, all the solutions of the state of the art that aims to reduce the amount of push of a numerical keypad key for the insertion of a letter or character, and this way also to reduce KSPC index, necessarily imply the use of at least one recorded dictionary in the memory of the cellular telephone. As the cellular phones are becoming smaller and cheaper, the size of the memory, and consequently its space of storage, is a critical factor to add the functions that are becoming more frequent and desired by the consumer, such as digital cameras, storage of images and video, storage of data etc.

Thus, it is desirous to free the space of storage in the memory of cellular phones, as much to allow the reduction of the size of the device, as to hold new functions and/or resources associated to the cellular phone. In addition, it is also advantageous to reduce the footprint associated to a functionality of a cellular telephone to permit a smaller space of storage in the memory.

Therefore, it is a first object of the present invention to provide a method of inputting one or more characters by means of a numerical keypad wherein it is possible to reduce the number of push of the said numerical keypad for said insertion of one or more characters, in comparison to the

Multitap technique.

It is another object of the present invention to provide a method of inputting one or more characters by means of a numerical keypad wherein the number of push of the keypad for the insertion of the said one or more character is equal to one.

Still another objective of the present invention is to provide a method of inputting one or more characters by means of a numerical keypad wherein it is possible to reduce the size of the memory, when compared with the known techniques as T9, iTap or eZiText.

DESCRIPTION OF THE INVENTION

The objects above among others are achieved with the present invention, which consists of a method of forming a word by inputting one or more characters by means of a numerical keypad that comprises the steps of:

(i) from a training lexicon, defining the contextual questions used to map the characters corresponding to the keys pushed in a numerical keypad;

(ii) to register each push of the key in the keypad that corresponds to the group of characters that contains the character that is desired to be inserted in a screen;

(iii) to analyze, at each push of the key, the context of the keypad keys pushed before and after each one of the pushed keys;

(iv) to show, at each push of the keypad key, the corresponding character of each pushed key on the basis of the contextual questions stored in memory, said characters being shown in a screen; and

(v) to finish the formation of the word when any corresponding keypad key corresponding to the punctuation, space or any special character is pushed. According to an embodiment of the invention, at least one character is inserted by means of a numerical keypad with only one push in said numerical keypad. In other words, for the insertion it is not necessary the relative number of push corresponding to the sequence of the character, but only one push in the corresponding keypad key to the character. For example, in the formation of the word "home", it is used the key containing the letter "h" in the corresponding keypad key to number "4" of the numerical keypad, and it is not necessary to push two times said keypad key, however only one time.

The same rule is applied in case it is desired to enter more than one character that is associated with the preceding character. For example, for the insertion of the word "home", it is necessary to push only the numerical keys "4", "6", "6" and "3", without repetition.

The first step (i) is executed only one time for each training lexicon, defining a set of contextual rules that is stored in memory and that are consulted at each push of the key in order to define the characters to be shown.

The training lexicon is the set of words or used phrases to create the contextual questions that, in turn, are the tests that form the set by means of which the ambiguities are reduced. The set of contextual rules can be stored in a not re- recordable memory at the moment of the manufacture of the devices that use the present method, or in re-recordable memory, allowing its update or change during the use of the related devices.

Later, (ii, iii) the method of the present invention is carried out by means of the analysis of the contained digits to the left (resultant of the previous push) of the character under analysis, and the analysis of the digits contained to the right (resultant of the posterior push) of the character under analysis, within an association of characters, according to the rules extracted of the construction of words of a specific language and vocabulary. This analysis is repeated for each keypad key to each push of the keypad key. The context of the digits to the right and the left allows, by means of the consultation to the rules, (iv) to define the character to be shown from the sequence of keypad keys pushed in the numerical keypad.

When a character of punctuation, space or any another special character (v) is inserted, the method is finished and the corresponding word is formed.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 represents a numerical keypad of a telephone, particularly of a cellular telephone; Figure 2 schematically represents the method of the present invention;

Figure 3 is a representation of the association of characters to the lexicon used in the method of the present invention; The figures 4a to 4c illustrate the analysis of the contexts to the right and the left during the insertion of the characters by means of a numerical keypad, in accordance with the method of the present invention; and

Figures 5, 6 and 7 illustrate a non-limiting example of the method according to the present invention. DESCRIPTION OF PARTICULAR EMBODIMENTS

Figure 1 represents a typical numerical keypad of a telephone, particularly of a cellular telephone. As it is known, these keypads comprise a series of letters associated to each numerical key. For example, the key of the number "2" has associated the letters "a", "b" and "c". The keypad key of the number "3" has associated the letters "d", "e" and T, and thus successively until the key of the number "9". Other keys, such as the corresponding key of the number "1", the key of the number "0", the key of the asterisk (*), and the pound (#) are destined to other functions, such as insertion of space between words, punctuation, symbols etc. The most common method of inserting characters from a numerical keypad, particularly the insertion of letters associated to numerical keys of a keypad, follows the logic of the sequence where they appear in the numerical keypad key. That is, for the insertion of the letter "o", the keypad key of the number "6" is necessary to be pushed three times continuously. In the case of being necessary the insertion of two subsequent letters contained in the same keypad key, as for example in the case of the word "hi", the key of the number "4" must be pushed two times, it is necessary to wait a moment, and the same key must be pushed again three times. O

According to the present invention, it is aimed to reduce the number of push in a numerical keypad for the formation of a word. Therefore, according to the example above, for the insertion of the word "hi" it is necessary to push subsequently and without the necessity of wait two times the keypad key of the number "4".

The method of the present invention is accomplished by analyzing the keypad keys pushed to the left of the character of reference that is being analyzed, and also analyzing the keypad keys pushed to the right of the character that is being analyzed, and thus successively from the first to the last character of the word to, in this manner, forming a word. This analysis is repeated for all the characters to each push of the keypad key, until the formation of the word is finished.

Figure 2 schematically illustrates the method of the present invention. First, a mapping of characters is related (10) for each keypad key of the numerical keypad. For example, for the keypad key of the number "2" the characters "a", "b" and "c" are associated, and thus successively until the key of the to number "9", as illustrated. Therefore, for each keypad key of the numerical keypad, there is a determined number of corresponding characters. Such structure can be summarized in the following formula:

N = Cp (I) where (N) is the corresponding keypad key to the determined number of the numerical keypad (N = 0, 1 , 2, 3...., 9); (C) is the character associated to the keypad key; and (p) is the amount of characters associated to the keypad key (N). For example, for the keypad key of the number "5" there are 3 characters associated to it, i.e. for N = 5, p = 3, where Ci = j, C₂ = k and C₃ = I.

An appropriate lexicon (20) for the used language, such as Portuguese, English, French etc, and for the type of the vocabulary, such as medical vocabulary, legal vocabulary, teenager vocabulary etc is also defined. It is generated then an association (30) between the mapped characters and the defined lexicon. Figure 3 illustrates such association. For example, for the word "cab", it is associated three times the keypad key of the number "2", because it contains all the characters that form the word. This association is carried out successively for all the vocables defined in the lexicon.

This step consists of the definition of a training lexicon. This way, and from the training lexicon, the definition of the contextual questions is carried out to reduce all the ambiguities between characters. This reduction of ambiguities is accomplished by constructing a number of contextual questions, divided in a series of standard questions (Q), wherein Q = q1 , q2 ..., qn, and being (q) each question formed from the number of standard questions (Q). The number of standard questions (q) is defined according to determined criteria, which are: - each question (q) refers only to a unique variable;

- the number of standard questions (Q) is related to the number of characters to the left (m) or the right (n) of the keypad key under analysis (N) in relation to which the context be will analyzed; and

- the relationship of the context with the amount of characters related to the lexicon (C1 , C2..., Cp).

In this in case, the number of standard questions is:

Q = (m+n)x8 (4)

Therefore, for each question (q), being q = 1 , 2..., Q; one determined question is adopted (q¹) so that, in one given situation, the contexts to the left (m) and the right (n) correspond to a unique character (d, C₂, ..., C_p). This adoption of the question (q¹) obeys the following relation: q' = arg max P(Cj | q), with i = 1 , 2, ..., P; and q = 1 , 2 Q (5) being that P (QDq) means the frequency of the conversions of a determined keypad key (N) under analysis for a character (Cj), i.e. it is taken into account the number of times that one determined keypad key (n) was converted for the character (Ci) and it is divided for the number of occurrences of the keypad key (N). The subdivision continues until that the original data set is divided in subgroups having no conversion ambiguity or having acceptable ambiguity. Therefore, it is obtained a structure with the best questions to convert a determined keypad key (N) into a character (Q). Variations of the formula (5) may use concepts of minimization of the entropy or minimization of the average square error. For the definition of the word to be inserted from the numerical keypad, as mentioned before, it is then carried out an analysis of the context of the keypad key used to the left of one determined keypad key (N) under analysis, and of keypad key used to the right of one determined keypad key under analysis. This analysis can be defined as: E^m(T) (2)

Dⁿ(T) (3) where (2) indicates that one determined key (T) occurred (m) times before or to the left of the keypad key (N) under analysis, and in (3) it means that one determined key (T) occurred (n) times after or to the right of the keypad key (N) under analysis. The figures 4a to 4c schematically illustrate the proposed relation. According to this example, for the word "text" is analyzed to the left and to the right, repeatedly and while the word is formed with the keypad keys that had been pushed. In other words, in the figure 4a it is firstly analyzed the context of the pushed keys to the right of the key of the number "8", and then the context of the pushed keys to the left of the key of the number "8", and successively in relation to the keypad keys of the numbers "3", "9" and "8", thus defining the word to be inserted, according to figure 4c.

The method of the invention is particularly applied in cellular phones for sending messages known by the acronym SMS. Other applications, such as sending of diverse SMS messages, storage and search of information in electronic agendas, remote controls of electro-electronic devices, such as television sets, DVD devices, Compact Disc devices, audio equipment, home theaters, personal computers etc, for the insertion of text, such as edition of the name of archives and directories, storage and search of information etc can also use the method of the present invention.

EXAMPLE

The present invention will be described in relation to a non-limiting example, without imposing any limitation beyond those defined in the attached claims.

It is taken as an analysis example the key of the number "9" that can represent four different characters in one same word. Figure 5 illustrates 4 different words that contain, in the structure, a corresponding keypad key to the number "9". These words are "text", "gauze", "axle", and "zero". In the column of the right, there is the numerical keypad keys that must be pushed to form the corresponding words represented in the column to the left.

According to this example, relations E^m(t) and Dⁿ(t) are applied in the following way: if D¹ (6), then use the letter "x", on the contrary, if E¹ (2), then to use letter "z", if not, use letter "x".

The method exemplified above may be represented according to the flowchart of figure 6.

For this example, the set of questions can be optimized in the following way: if E¹ (2), then use the letter "z", on the contrary use the letter "x". The method above exemplified also can be represented in a flowchart according to the one of figure 7.

As it was described specifically for the corresponding keypad key to number "9", the method of the present invention can be expanded for all the keypad keys of the numerical keypad, thus forming a structure for each keypad key and reducing the ambiguities in the system of formation of inserted words from a numerical keypad.

The amount of questions and the number of necessary contexts to the left and to the right in order to reduce the ambiguities may vary depending on the size and the variety of the used lexicon.

Evidently, and according to the above specification, the method of the invention reduces the footprint in relation to the methods of the state of the art, beyond allowing an adjustable footprint depending on the requirements or the necessities of applications. This is because the method of the present invention eliminates the necessity of using a dictionary with sets of thousands of words, being stored only the set of contextual questions of the contexts to the left and to the right for each one of the keypad keys. Thus, the necessary space in the memory is also reduced to take the effect the insertion of words from a numerical keypad. Finally, due to the proper nature of the method of the invention, the KSPC index is reduced to one.

It must be recognized that the present invention has been described with regard to a particular embodiment. Those skilled in the art may develop a wide variety of details and expand the above-described method for other types of applications without, however, deviating from the principles of the invention. Therefore, the attached claims must be interpreted as covering all the equivalents that are in the scope and character of the invention.

Claims

1. METHOD OF FORMATION OF A WORD BY INPUTTING CHARACTERS BY MEANS OF A NUMERICAL KEYPAD, characterized by comprising the steps of: (i) from a training lexicon, defining the contextual questions used to map the characters corresponding to the keys pushed in a numerical keypad;

(ii) to register each push of the key in the keypad that corresponds to the group of characters that contains the character that is desired to be inserted in a screen; (iii) to analyze, at each push of the key, the context of the keypad keys pushed before and after each one of the pushed keys;

(iv) to show, at each push of the keypad key, the corresponding character of each pushed key on the basis of the contextual questions stored in memory, said characters being shown in a screen; and (v) to finish the formation of the word when any corresponding keypad key corresponding to the punctuation, space or any special character is pushed.

2. The Method of claim 1 , characterized in that said analysis is carried through the relation E^m(T) and Dⁿ(T), where (T) is an occurring key, (m) and (n) indicate the position where said key (T) occurred to the left (E) and the right (D) of the character under analysis, respectively.

3. The Method of claim 1 , characterized in that said lexicon of training and said mapping of characters of the said numerical keypad keys are related according to the relation N = C_p; where N is a determined keypad key of the numerical keypad; (C) is a determined character associated with the said key; and (p) is the amount of characters associated to the said keypad key (N).

4. The Method of claims 1 or 3, characterized in that an association between the said lexicon and the said mapping of characters is carried out.

5. The Method of claim 1 , characterized in that said training lexicon is defined in accordance with the language and application of the method for areas of specific vocabulary.

6. The Method of claim 1, characterized in that the contextual questions are defined from a number of standard questions (Q) that obeys the relation Q = (m+n)x8, where (m) is the amount of characters to the left (m) and (n) is the amount of characters to the right in relation to a key (N) under analysis. 7. The Method of claim 1 or 6, characterized in that said contextual questions are obtained from a set of questions defined by the relation q' = arg max P(Ci I q), with i = 1 , 2, ..., P; and q = 1 , 2, ..., Q; where P (CjDq) is the frequency of the conversions of a determined keypad key (N) under analysis for a character (Cj), and (q¹) is a question raised from the highest probability.

8. Use of the method as defined in one of claims 1 to 7 in cellular telephones for text insertion.

7. Use of the method as defined in one of claims 1 to 7 in remote controls of electro-electronic devices.