JP2011090356A - Data processor, computer program thereof, and data processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data processor for excellently preventing an erroneous input even if an input means such as a touch panel being an input device is compact in size. <P>SOLUTION: A plurality of operation keys of the input means 11 such as a touch panel being an input device are selectively and continuously inputted by operations, but a calculating means 12 individually calculates appropriate probabilities based on respective prescribed physical amounts when the operation keys are substantially simultaneously inputted by operations. Since a model storage means 16 stores a statistical model corresponding to the linkage probability of the operation keys continuously inputted by operations, a search means 13 detects the statistical model based on the linkage probability with the operation key which is inputted by an immediately preceding operation for each of the operation keys which are substantially simultaneously inputted by operations. Then, prescribed arithmetic processing is performed by the use of the appropriate probability and the statistical model, so that an output means 17 selects one of the operation keys which are substantially simultaneously inputted by operations. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a data processing apparatus using a touch panel as an input device, and more particularly to a data processing apparatus that prevents erroneous operation of a small touch panel, a computer program thereof, and a data processing method.

  As an input means to the computer device, there is a touch input in a terminal having a touch panel. The input using the touch panel can change the contents of the displayed keys, and is excellent as an intuitive operation of the user with respect to the computer.

  On the other hand, when the screen is small, such as a portable terminal, erroneous input such as touching a key that is not the target key by mistake or touching a plurality of keys at the same time becomes a problem. In particular, when a touch panel is used as an input device for a mobile phone or PDA (Personal Digital Assistance) that does not have a keyboard, a software keyboard may be used. However, if the screen is small, there is a high possibility of erroneous input. Therefore, there is a risk of hindering the input of characters and words.

  In order to solve such problems, there is a proposal to reduce erroneous input by considering probability values even when multiple keys are touched by using statistical information such as word dictionary and character chain (For example, refer to Patent Document 1).

  There is also a proposal for easily and accurately obtaining position information used for alignment with predetermined points of a plurality of partitioned areas on an object. In the proposal, a plurality of combinations of information of k temporary abnormal values and (n−k) temporary normal values are extracted from n pieces of actually measured position information obtained by measuring a plurality of marks.

  Next, for each statistical model corresponding to each extracted combination, a predetermined model parameter and AIC (An Information Criterion) are calculated, and the probability of the model based on the AIC is estimated.

  Next, based on the estimation results, a statistical model that is most likely to be selected is selected from all statistical models, and the model parameters of the selected statistical model are aligned with predetermined points of a plurality of partitioned areas on the object. It is determined as the statistically most appropriate value of the model parameter of the model formula for calculating the position information used for (see, for example, Patent Document 2).

  It also records and learns the touch position of the user's touch panel by estimating the physical features such as how to hold the remote control equipped with the touch panel, finger size, physical characteristics such as the dominant hand, and how to push. Thus, there is also a proposal of an input device and method for automatically correcting an error between a contact position and a button position on a regular software keyboard.

  In the proposal, in an input device that performs input by pressing one or a plurality of buttons displayed on the touch panel, a touch detection area defining unit that defines a touch detection area on the touch panel for each button; A history recording unit that records input information, a first determination unit that determines whether the contact position of the user is included in any of the contact detection areas of each button, and the first determination unit that is not included in the first determination, In the history record information, a second determination means for determining whether the contact position is included in any of the contact detection areas, and if included in the second determination, the contact position is added to the determined contact detection area Position adding means (see, for example, Patent Document 3).

  Furthermore, in the touch panel display device, there is also a proposal of a touch panel display device that can reduce not only a visual field shift but also an erroneous input due to vibration or acceleration. In the proposal, the displacement amount (speed, acceleration, moving direction, angle) of the display unit of the touch panel display device is detected by a sensor, and the recognition area on the touch panel is corrected by the CPU based on the displacement amount.

  At this time, the recognition area is corrected to move at a speed slower than the moving speed of the display unit, or the recognition area is corrected to a trapezoid shape at a speed slower than the tilting speed (see, for example, Patent Document 4).

US Patent Publication No. 2006/0085757 Japanese Patent Laid-Open No. 2004-006332 JP 2008-242958 A JP 2009-009261 A

  However, with the conventional method, it is difficult to prevent erroneous input well when the input device is a small touch panel or the like.

  The present invention has been made in view of the problems as described above, and provides a data processing apparatus, a computer program thereof, and a data processing method capable of satisfactorily preventing erroneous input even when the touch panel is small.

  The data processing apparatus of the present invention is a data processing apparatus that detects a selective and continuous input operation of a plurality of operation keys of an input device, and each of the predetermined operations when the plurality of operation keys are input at substantially the same time. A search means for searching a statistical model based on a chain probability between a calculation means for individually calculating an appropriate probability from a physical quantity of each and a plurality of operation keys that have been input at approximately the same time and an operation key that has been input immediately before; And an output means for selecting one of a plurality of operation keys that are input at substantially the same time by executing predetermined arithmetic processing using the statistical model.

  The computer program of the present invention is a computer program for a data processing device that detects a selective and continuous input operation of a plurality of operation keys of an input device, and each time when the plurality of operation keys are input at substantially the same time, Statistical search processing for searching a statistical model from the chain probability of the operation key input immediately before each of a plurality of operation keys input at substantially the same time. Then, the data processing device is caused to execute an operation selection process for selecting one of a plurality of operation keys that are input at substantially the same time by executing a predetermined calculation process with the appropriate probability and the statistical model.

  The data processing method of the present invention is a data processing method for a data processing apparatus for detecting a selective and continuous input operation of a plurality of operation keys of an input device, when the plurality of operation keys are operated to be input substantially simultaneously. Statistics that search for a statistical model from the chain probability of the appropriate calculation operation that calculates the appropriate probability individually from each predetermined physical quantity and the operation key that was input immediately before each of the operation keys that were input almost simultaneously A search operation, and an operation selection operation for selecting one of a plurality of operation keys that are input at substantially the same time by executing a predetermined arithmetic process using the appropriate probability and the statistical model.

  It should be noted that the various components of the present invention need only be formed so as to realize their functions. For example, dedicated hardware that exhibits a predetermined function, data processing in which a predetermined function is provided by a computer program It can be realized as an apparatus, a predetermined function realized in the data processing apparatus by a computer program, an arbitrary combination thereof, or the like.

  The various components of the present invention do not necessarily have to be independent of each other. A plurality of components are formed as a single member, and a single component is formed of a plurality of members. It may be that a certain component is a part of another component, a part of a certain component overlaps with a part of another component, or the like.

  Moreover, although the computer program and the data processing method of this invention have described several process and operation | movement in order, the order of description does not limit the order which performs several process and several operation | movement.

  For this reason, when implementing the computer program and data processing method of the present invention, the order of the plurality of processes and the plurality of operations can be changed within a range that does not hinder the contents.

  Furthermore, the computer program and the data processing method of the present invention are not limited to being executed at a timing when a plurality of processes and a plurality of operations are individually different. For this reason, other processes and operations occur during execution of certain processes and operations, and the execution timing of certain processes and operations overlaps with the execution timing of other processes and operations. Etc.

  The data processing apparatus according to the present invention can read a computer program and execute a corresponding processing operation, so that a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), an I / F It can be implemented as hardware constructed by general-purpose devices such as (Interface) units, dedicated logic circuits constructed to execute predetermined processing operations, combinations thereof, and the like.

  In the present invention, causing the data processing apparatus to execute various operations corresponding to the computer program also means causing the data processing apparatus to control operations of the various devices.

  For example, storing various data in the data processing device means that the CPU stores various data in an information storage medium such as an HDD (Hard Disc Drive) fixed to the data processing device, and can be exchanged with the data processing device. The CPU allows various data to be stored by the CD drive in an information storage medium such as a CD-R (Compact Disc-Recordable) loaded.

  Furthermore, “store data” in the present invention means that the apparatus of the present invention has at least a function of storing data. For this reason, in addition to storing data newly registered by the consumer, the device of the present invention also stores data registered at the time of manufacture by the supplier, and allows new data not to be registered by the consumer after shipment. .

  The “appropriate probability” referred to in the present invention is also called “likelihood” or the like, and is a probability that the phenomenon occurs under a certain condition. In the present invention, for example, there is a certainty that a specific operation key has been input when a certain physical quantity is observed.

  Furthermore, the “chain probability” as used in the present invention is the probability that another (or the same) phenomenon B occurs after a certain phenomenon A. A “statistical model” is a model that is statistically learned from a large amount of data.

  In the data processing apparatus according to the present invention, a plurality of operation keys of the input device are selectively operated continuously. However, when a plurality of operation keys are input at substantially the same time, the calculation means individually calculates the appropriate probability from each predetermined physical quantity. In addition, since the model storage unit stores a statistical model corresponding to the chain probability of the operation keys that are continuously input, the operation key that is input immediately before each of the plurality of operation keys that are input almost simultaneously. The search means detects a statistical model from the chain probability. Then, a predetermined calculation process is executed with the appropriate probability and the statistical model, and one output means is selected from a plurality of operation keys input at substantially the same time. For this reason, for example, erroneous input can be satisfactorily prevented even when a small touch panel is used as an input device.

It is a block diagram which shows the structure of the data processor concerning embodiment of this invention. It is a flowchart which shows the input processing concerning embodiment of this invention. It is a block diagram which shows the whole structure of the character input system concerning Example 1 of this invention. It is a flowchart which shows the character input process concerning Example 1 of this invention. It is an example of the input process concerning Example 1 of this invention. It is a block diagram which shows the structure of the data processor concerning the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the data processor concerning the 2nd Embodiment of this invention. It is a block diagram which shows the whole structure of the address search system concerning Example 2 of this invention. It is a flowchart which shows the address search process concerning Example 2 of this invention.

  An embodiment of the present invention will be described below with reference to the drawings. The data processing apparatus 10 according to the present embodiment detects selective and continuous input operations of a plurality of operation keys of the input unit 11 such as a touch panel as an input device.

  For this reason, as shown in FIG. 1, the data processing apparatus 10 according to the present embodiment has a calculation unit 12 that individually calculates appropriate probabilities from respective predetermined physical quantities when a plurality of operation keys are input at substantially the same time. A chain of model storage means 16 that stores a statistical model corresponding to the chaining probability of operation keys that are continuously input and an operation key that is input immediately before a plurality of operation keys that are input approximately simultaneously Search means 13 for searching for a statistical model from the probability, and output means 17 for selecting one of a plurality of operation keys that are input at substantially the same time by executing predetermined arithmetic processing with the appropriate probability and the statistical model. .

  More specifically, the data processing apparatus 10 according to the present embodiment is, for example, a general-purpose computer system, and includes a CPU, RAM, ROM, and a nonvolatile storage device such as an HDD as components not shown. .

  In the data processing apparatus 10 according to the present embodiment, the CPU reads an OS (Operation System) stored in a RAM, a ROM, or a non-volatile storage device and a computer program of an application, and executes an input process.

  Such a computer program includes, for example, a chain of a proper calculation process that individually calculates a proper probability from each predetermined physical quantity when a plurality of operation keys are input at substantially the same time, and an operation key that is continuously input. A model storage process for storing a statistical model corresponding to a probability, a statistical search process for searching a statistical model from a chain probability of an operation key input immediately before each of a plurality of operation keys input at substantially the same time, and appropriate It is described to cause the data processing apparatus 10 to execute an operation selection process that executes a predetermined calculation process with the probability and the statistical model and selects one from a plurality of operation keys that are input at substantially the same time.

  Thereby, when the operation key on the input means 11 which is a touch panel is touched as an input operation, the touched operation key can be output with high accuracy. Note that the data processing apparatus 10 does not have to be a single computer system, and may be configured by a plurality of computer systems.

  The calculation means 12 obtains the probability of input of each operation key for a plurality of operation keys that are touched candidates from the physical quantity at the time of input. Specifically, for example, when an occupied area is used as a physical quantity, the area of each operation key when a plurality of operation keys are touched is obtained, and this is output as the probability of each operation key, that is, the distance value. For example, the distance value Dist (A) of the operation key “A” can be determined to be closer to the operation key having a larger occupied area, that is, more likely by using the reciprocal of the occupied area S (A).

  Here, the occupation area is used as the physical quantity at the time of input, but the physical quantity is not limited to this. For example, the pressure at the time of input, the speed or movement amount when the operation keys are continuously input, or the like may be used.

  When the pressure at the time of input is used, a pressure sensor may be provided on the touch panel, the pressures at a plurality of operation keys may be measured and used as distance values. As a result, for example, an operation key having a higher pressure can be considered to be more reliable.

  When using the speed when continuously inputting an operation key, the speed between touches of the operation key is obtained by measuring the time from when a certain operation key is pressed until the next operation key is pressed, This may be used as a physical quantity considering the context information called operation key chain.

  Similarly, when using the movement amount when inputting an operation key continuously, by measuring the distance from one operation key being pressed to the next operation key being pressed, What is necessary is just to obtain | require a movement amount and to use this as a physical quantity.

  Thereby, for example, the certainty when a certain operation key is pressed and then the next operation key is pressed is more probable that the speed or movement amount is closer to, for example, a separately determined standard value (average value). can do.

  The search means 13 uses the statistical model stored in the model storage means 16 to search for the optimum operation key using the distance value of each operation key as the output of the calculation means 12 as an input. More specifically, for example, assuming a Japanese sentence input, and using a character Ngram that is a character chain model as a statistical model, the character Ngram is learned by using, for example, ease of connection of characters in Japanese as a model. The probability of each input operation key is obtained from the distance value output from the calculation means 12 and the character chain probability, and the most probable operation key is output.

  Here, the character chain model is used as the statistical model, but the present invention is not limited to this. For example, assuming a continuous operation key input operation to realize a certain function, such as operating a video device on the touch panel, the operation can be performed by using a model that has learned several operation key chains. It is possible to improve the accuracy of key input operations. In this case, the learning of the chain model of the operation keys may be dependent on the device or function, or may be learned depending on the user to be used.

  In addition, although the data processing apparatus on a touch panel was mentioned as an example here, if a physical quantity can be measured at the time of input, it will not be restricted to this. For example, the same applies to an apparatus that selects and inputs an operation key displayed on a TV screen or the like with a pointing device.

  The statistical model may be a non-volatile storage device such as a hard disk drive, a magneto-optical disk drive, or a flash memory, or may be a volatile storage device such as a DRAM (Dynamic Random Access Memory). The statistical model may be a storage device externally attached to the data processing device 10.

<Operation of Best Mode 1>
Next, the input processing according to the embodiment of the present invention will be described with reference to the flowchart shown in FIG. First, the data processing apparatus 10 accepts an operation key input operation (S100). Specifically, the touch of the operation key is detected by the touch panel.

  Next, the data processing apparatus 10 performs distance calculation (S101). Specifically, for example, when using the occupied area as a physical quantity at the time of input, the calculation unit 12 of the data processing device 10 calculates the area S of the touched region for a plurality of touched operation keys. Each is obtained and the reciprocal of this is output as a distance value Dist.

  Next, the data processing device 10 performs an operation key search (S102). Specifically, the search means 13 of the data processing apparatus 10 assumes, for example, a Japanese sentence input, and uses a character Ngram that is a character chain model as a statistical model. Learn how easy it is to connect.

Input from the distance value Dist (A) of the operation key “A” as the output of the calculation means 12 and the probability P (h | A) that the operation key “A” is chained next to the character chain h up to that point The probability P (A) that the assigned operation key is "A"
P (A) = Dist (A) × P (h | A)
Asking.

  By calculating and comparing the probabilities of the operation keys input in this way, the most probable, that is, the operation keys with the highest probability are output as the search results. Further, an output means 17 may be provided to output the search result obtained in step S102 (S103). The output means may separately prepare a screen as a display device, for example, or may display on a touch panel.

<Effect of Best Mode 1>
As described above, the data processing apparatus 10 according to the present invention outputs the most probable operation key using the physical quantity and the statistical model at the time of input when a plurality of operation keys are input on the touch panel, that is, erroneous input. Can be reduced.

  A method that uses only a statistical model such as a character chain is a character chain specialized for Japanese input when you want to input a character chain that is rare in Japanese input (for example, foreign words). However, this model has a problem in that an erroneous input occurs due to adverse effects, and the target input becomes difficult.

  Even in such a case, the present invention considers a physical quantity at the time of input, that is, a physical quantity such as an occupied area. For example, when the occupied area of the target operation key is extremely large, the influence of the character chain model is relatively It can be reduced, and it can be said that such problems are less likely to occur.

<Example 1>
An example of a character input system will be described below as Example 1 of the data processing apparatus according to the present embodiment. FIG. 3 is a diagram illustrating the overall configuration of the character input system according to the first embodiment. A character input system 100 that is a data processing apparatus shown in FIG. 3 includes an input unit 110, an input processing unit 10b, and an output unit 120.

  The character input system 100 is a general-purpose computer system, and includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and a nonvolatile storage device as components not shown. In the character input system 100, the CPU reads an OS (Operation System) and a character input program stored in a RAM, a ROM, or a nonvolatile storage device, and executes a character input process.

  Thereby, the character input using a touch panel etc. can be performed accurately. Note that the character input system 100 does not have to be a single computer system, and may be composed of a plurality of computer systems.

  The input means 110 is a touch panel provided with a touch sensor, for example. The input processing unit 10b corresponds to the data processing device 10 of FIG. Therefore, in the following, the difference from FIG. 1 will be mainly described, and the description of the configuration corresponding to FIG. 1 and having the same function will be omitted.

  The calculation means 12 obtains the certainty from the physical quantity at the time of input of each operation key input by the input means 110, and outputs it as a distance value. The search means 13 obtains the probability of input of each operation key from the distance value using a statistical model, searches for the most probable operation key, and outputs the result. The output unit 120 is, for example, the same touch panel as the input unit 110, and displays the operation keys as a result of the search unit 13 on the touch panel.

<Operation of Example 1>
Next, the overall flow of input processing according to the first embodiment of the present invention will be described with reference to the flowchart shown in FIG. First, the character input system 100 accepts an input operation of an operation key (S200).

  Specifically, the character input system 100 receives an input of an operation key touched on a touch panel as the input unit 110. If there is no input (S201), the character input process is terminated.

  If there is an input, the character input system 100 executes an input process (S202). Specifically, input processing as shown in FIG. 2 is performed by the calculation means 12 and the search means 13 in the input processing unit 10b of the character input system 100, and search results are output.

  For example, as shown in FIG. 5, when the word “voice” is entered on the software operation keyboard and “o”, “n”, “s” are entered, If the input is an area as shown in FIG. 5, the area of the operation keys “e”, “w”, “3”, etc. is calculated as a distance value.

Also, considering the probability of chaining with the previous input "s", the probability of each operation key is
P (e) = 0.8 × 0.20 = 0.160
P (w) = 0.8 × 0.02 = 0.016
P (3) = 0.2 × 0.10 = 0.020
As a result, the operation key “e” is output from P (e) having the highest probability value, that is, the probability value.

  Next, the character input system 100 outputs the search result obtained in the input process (S203). Specifically, the character obtained as a result of the input process in step S202 is displayed on the display device by the output means 120 of the character input system 100.

  The display device may be the same touch panel as the input unit 110. After outputting the result, the process returns to step S200 to accept the next input. As described above, in the first embodiment of the present invention, it is possible to continuously display operation keys input from the touch panel for inputting characters.

<Best Mode 2>
In the second embodiment of the present invention, compared to the first embodiment, a touch panel is used by providing a feature quantity model statistically learned using a feature vector whose feature quantity is a physical quantity at the time of input. This improves the accuracy of input.

  FIG. 6 is a diagram showing an overall configuration of the data processing apparatus according to the second embodiment. When the operation key on the touch panel is touched, the data processing device 10c in FIG. 6 outputs the touched operation key.

  The data processing apparatus 10c according to the present embodiment further includes a model storage unit 14 that stores a feature quantity model that is statistically set using a feature vector having a characteristic value as a feature quantity. Then, the appropriate probability is detected by calculating the distance value between the detected characteristic value and the feature amount model.

  More specifically, the data processing device 10c of the present embodiment is a general-purpose computer system, and includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and a non-volatile memory as configurations not shown. Equipment. In the data processing device 10c, a CPU reads an OS (Operation System) and an input processing program stored in a RAM, a ROM, or a nonvolatile storage device, and executes input processing.

  Thereby, when the operation key on the touch panel is input, the input operation key can be output with high accuracy. The data processing apparatus 10c does not have to be a single computer system, and may be configured by a plurality of computer systems.

  As shown in FIG. 6, the data processing apparatus 10c of the present embodiment includes a calculation unit 12c, a search unit 13, a feature amount model, and a statistical model. Here, the search means 13 and the statistical model are the same as those in FIG. Only the differences from FIG. 1 will be described below.

  The calculating means 12c calculates the distance from each operation key for a plurality of operation keys that are input candidates from the physical quantity at the time of input using the feature amount model. Specifically, for example, when an occupied area is used as a physical quantity at the time of input and the feature model is learned using a feature vector having the occupied area of each operation key as a feature quantity, a plurality of operation keys are input. Then, the area S of each operation key is calculated, and the distance from the feature quantity model representing each operation key is calculated using this as the feature vector X, and the distance value Dist of each operation key is output.

  The technology for statistically learning a model from a large amount of feature vectors, and the technology for calculating the distance between the feature vector and the model to be input and outputting the distance value are well known in the field of speech recognition, etc. The description is omitted here. As in FIG. 1, the search means 13 uses the statistical model to search for the optimum operation key using the distance value of each operation key as the output of the calculation means 12 as an input.

  As shown in FIG. 7, the data processing device 10e further includes a word dictionary 15 in which a plurality of word data consisting of a chain of input operations of a plurality of operation keys is registered. A statistical model may be detected corresponding to word data registered in the word dictionary. In this case, the search means 13c searches for a word included in the word dictionary 15 to improve accuracy by specializing in the output of the word.

  Further, the word dictionary 15 is set with a second statistical model corresponding to the chaining frequency of the input operation of the operation key for each of a plurality of registered word data. The word data may be searched for corresponding to.

  In this case, it is possible to improve the accuracy of word output. A technique for increasing the accuracy of word output using a word dictionary or a statistical model representing a word chain is a known technique in the field of speech recognition and the like, and thus the description thereof is omitted here.

<Operation of Best Mode 2>
Next, audio processing according to the second embodiment of the present invention will be described. Here, the second embodiment of the present invention is the same as FIG. 2 except for the distance calculation processing in step S101 in FIG. In step S101, the data processing device 10c performs distance calculation using the feature amount model.

  Specifically, the calculation unit 12c of the data processing device 10c uses, for example, an occupation area as a physical quantity at the time of input, and the feature quantity model is learned using a feature vector having the occupation area of each operation key as a feature quantity. In this case, first, an area S of each operation key when a plurality of operation keys is input is obtained, and this is used as a feature vector X.

That is, the vector X having the area of the operation key “A” as S (A) and the area of each operation key as one dimension.
X = {S (Q), S (W), S (E), S (R), S (T), S (Y), ...}
Ask for.

  The dimension number of the vector X is the number of operation keys that can be input, that is, the number of operation keys displayed on the touch panel. If the feature model is learned for each operation key from a large amount of learning data similar to the feature vector X, the distance value Dist (A) of the operation key “A” is the model M_A representing the operation key “A”. What is necessary is just to obtain | require as a distance with the vector X. The distance value for each operation key thus obtained is used as the output of the calculation means 12c.

  Note that the feature amount is expressed as a feature vector having the area of each operation key as one dimension, but is not limited to the above-described method. For example, the area of the touch panel may be considered as one image as in image recognition and may be represented as a two-dimensional feature amount.

<Effect of Best Mode 2>
As described above, the data processing device 10c according to the second embodiment of the present invention further increases the accuracy of input using the touch panel by including the feature amount model. This is because a higher accuracy can be expected when the distance calculation is performed using the feature vector considering the physical quantities of all the operation keys that can be input than simply considering the physical quantity of one operation key. For example, it is possible to learn the relationship between a target operation key and an operation key that is easily mistaken from the arrangement of each operation key on the touch panel.

<Example 2>
An example of an address search system will be described below as Example 2 of the data processing apparatus according to the second embodiment of the present invention. FIG. 8 is a diagram illustrating the overall configuration of the address search system according to the second embodiment. The address search system 200 in FIG. 8 includes a display unit 210, an input unit 220, an input processing unit 10d, a changing unit 230, and an address database 240.

  The address search system 200 is a general-purpose computer system, and includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and a nonvolatile storage device as components not shown.

  In the address search system 200, the CPU reads an OS (Operation System) and an address search program stored in a RAM, a ROM, or a nonvolatile storage device, and executes an address search process.

  Thereby, the input of the address using a touch panel etc. can be performed with high precision. Note that the address search system 200 does not have to be a single computer system, and may be configured by a plurality of computer systems.

  The display unit 210 displays address candidates to be searched for, for example, prefecture names and city names according to the group according to the hierarchy or region. The input means 220 is a touch panel provided with a touch sensor, for example.

  The display unit 210 and the input unit 220 are preferably the same touch panel. Specifically, for example, the prefecture name displayed on the display unit 210 may be an operation key of the input unit 220.

  The input processing unit 10d corresponds to the data processing device 10c in FIG. Therefore, in the following, the difference from FIG. 6 will be mainly described, and the description of the configuration corresponding to FIG. 6 and having the same function will be omitted.

  The calculation unit 12c uses the feature amount model, obtains the probability of input of each operation key from the physical quantity at the time of input of each operation key input by the input unit 220, and outputs it as a distance value.

  The search means 13 obtains the probability of input of each operation key from the distance value using a statistical model, searches for the most probable operation key, and outputs the result. The changing unit 230 uses the address database 240 to narrow down addresses to be searched based on the search result obtained from the input processing unit 10d.

  Specifically, for example, the lower hierarchy of the address corresponding to the search result in the address database 240 is set as the next search target and fed back to the display unit 210 for the next input. For example, if the search result at the prefecture level is “Kanagawa Prefecture”, the names of cities, towns and villages under Kanagawa Prefecture, such as “Yokohama City”, “Kawasaki City”, “Sagamihara City”, etc. will be searched. .

  At this time, not only the address candidates to be displayed on the display unit 210 but also the feature amount model and the statistical model may be switched according to the address candidates to be changed.

  Note that the feature amount model and the statistical model may be adapted according to the use history of the user. Since the technique for adapting the statistical model to the user is a known technique in the field of speech recognition or the like, the description thereof is omitted here. In the present embodiment, the operation key to be input is a prefecture name or the like, but the same applies to the case where a map is displayed and touched on the map.

<Operation of Example 2>
Next, the overall flow of address search processing according to the second embodiment of the present invention will be described with reference to the flowchart shown in FIG. First, the address search system 200 displays the search target on the display unit 210 as an operation key for input (S300).

  Next, the address search system 200 accepts input operation keys that are search targets (S301). Specifically, the address search system 200 accepts input of operation keys input on a touch panel as the input unit 220.

  When the end is input (S302), the address search process ends. When an operation key other than the end key is input, the address search system 200 executes input processing (S303). Specifically, input processing is performed by the calculation unit 12c and the search unit 13 in the input processing unit 10d of the address search system 200, and a search result is output.

  Next, the address search system 200 executes a change process based on the search result obtained in the input process (S304). Specifically, the changing unit 230 of the address search system 200 uses the address database 240 to narrow down the search target addresses based on the search result obtained from the input processing unit 10d. The result of narrowing down is reflected in the display means 210, the feature amount model, and the statistical model.

  After the change process, the process returns to step S300 to accept the next input. As described above, in the second embodiment of the present invention, even when there are many operation keys to be input as in address search, it is possible to input with high accuracy by appropriately using the feature amount model.

  The present invention is not limited to the present embodiment, and various modifications are allowed without departing from the scope of the present invention. For example, in the above embodiment, the feature amount model is prepared in advance in the data processing apparatus 10.

  However, the data processing apparatus further includes: a history storage unit that stores an input history of the input unit 11 such as a touch panel that is an input device; and a model generation unit that statistically generates a feature amount model from the stored input history. You may have (not shown). In this case, since the feature amount model is generated from the input history of the user, it is possible to automatically set a feature amount model appropriate for the user's input operation.

  The model generation means as described above may statistically generate a feature amount model from an input history of a specific user, or may statistically generate a feature amount model from input histories of a plurality of users.

  When a feature amount model is generated from an input history of a specific user, the data processing device is specialized for that user. In addition, when a feature amount model is generated from input histories of a plurality of users, the data processing apparatus has good versatility.

  Further, for example, when an operation key chain is used as a statistical model, the key chain model is learned from an operation frequently performed by a certain user, and is calculated from a distance value that is an output of the calculation means 12 and an operation chain probability. The probability of each input operation key may be obtained, and the most probable operation key may be output.

  Needless to say, the above-described embodiments and examples can be combined within a range in which the contents do not conflict with each other.

DESCRIPTION OF SYMBOLS 10 Data processing apparatus 10b Input processing part 10c Data processing apparatus 10d Input processing part 11 Input means 12 Calculation means 12c Calculation means 13 Search means 14 Model storage means 15 Word dictionary 16 Model storage means 17 Output means 100 Character input system 110 Input means 120 Output means 200 Address search system 210 Display means 220 Input means 230 Change means 240 Address database

Claims (10)

A data processing device for detecting a selective and continuous input operation of a plurality of operation keys of an input device,
A calculation means for individually calculating a proper probability from each predetermined physical quantity when a plurality of the operation keys are input at substantially the same time;
Model storage means for storing a statistical model corresponding to the chain probability of the operation keys that are continuously input and operated,
Search means for searching the statistical model from the chain probabilities with the operation key input immediately before each of the plurality of operation keys input at substantially the same time;
An output means for selecting one of the plurality of operation keys that are input at substantially the same time by executing predetermined arithmetic processing with the appropriate probability and the statistical model;
A data processing apparatus. Model storage means for storing a feature quantity model statistically set using a feature vector having a characteristic value as a feature quantity;
The data processing apparatus according to claim 1, wherein the calculation unit detects the appropriate probability by calculating a distance value between the detected characteristic value and the feature amount model. History storage means for storing an input history of the input device;
The data processing apparatus according to claim 2, further comprising model generation means for statistically generating the feature amount model from the stored input history.   The data processing apparatus according to claim 3, wherein the model generation unit statistically generates the feature amount model from the input history of a specific user.   The data processing apparatus according to claim 3, wherein the model generation unit statistically generates the feature amount model from the input histories of a plurality of users. A word dictionary in which a plurality of word data composed of a chain of input operations of the plurality of operation keys is registered;
6. The data processing apparatus according to claim 1, wherein the search unit detects the statistical model corresponding to the word data registered in the word dictionary. In the word dictionary, a second statistical model corresponding to the chain frequency of the input operation of the operation key is set for each of the plurality of registered word data,
The data processing apparatus according to claim 6, wherein the search means searches for the word data corresponding to the second statistical model.   The calculation means calculates the appropriate probability using at least one of an occupied area where the operation key is input, a pressure of the input operation, a speed of the continuous input operation, and a movement amount of the continuous input operation as the physical quantity. The data processing device according to any one of claims 1 to 7. A computer program for a data processing device for detecting a selective and continuous input operation of a plurality of operation keys of an input device,
A proper calculation process for individually calculating a proper probability from each predetermined physical quantity when a plurality of the operation keys are input at substantially the same time;
A model storage process for storing a statistical model corresponding to the chaining probability of the operation keys that are successively input;
A statistical search process for searching a statistical model from a chain probability with the operation key input immediately before each of the plurality of operation keys input at substantially the same time;
An operation selection process for selecting one of the plurality of operation keys that are input at substantially the same time by executing a predetermined calculation process with the appropriate probability and the statistical model;
Is a computer program that causes a data processing apparatus to execute. A data processing method of a data processing apparatus for detecting a selective and continuous input operation of a plurality of operation keys of an input device,
Appropriate calculation operation for calculating appropriate probabilities individually from predetermined physical quantities when a plurality of the operation keys are input at substantially the same time;
A model storage operation for storing a statistical model corresponding to the chaining probability of the operation keys that are continuously input;
A statistical search operation for searching a statistical model from a chain probability with the operation key input immediately before each of the plurality of operation keys input at substantially the same time;
An operation selection operation for selecting one of the plurality of operation keys that are input at substantially the same time by executing predetermined arithmetic processing with the appropriate probability and the statistical model;
A data processing method.

Images