CN110837305A - Input method error correction method and device - Google Patents

Abstract

The invention discloses an input method error correction method and a device, wherein the method comprises the following steps: determining a holding posture type corresponding to a current key in a user input process; acquiring an error correction mapping relation table corresponding to the holding posture type; and carrying out error correction processing according to the error correction mapping relation table to obtain one or more input strings after error correction. The invention can improve the error correction capability and accuracy of the input method and improve the input quality of the user.

Description

Input method error correction method and device

technical field

The invention relates to the field of input method error correction, in particular to an input method error correction method and device.

Background technique

Input method is a coding method for inputting various symbols into computers or other devices. It is an indispensable tool for human beings to deal with computers. For example, Pinyin input is one of many input methods. Yes, no special memory is required, it is in line with human thinking habits, as long as you know Pinyin, you can input Chinese characters.

In the process of user input, sometimes the wrong button is clicked accidentally or the button is clicked at a fast speed, resulting in input error. Automatic error correction in the middle and background. In the prior art, the error correction technology of the input method is roughly divided into two types: button error correction and position error correction. Among them, the button error correction does not consider the positional relationship between the buttons, but only considers the logical relationship of the button sequence, such as: zle, button error correction may be corrected to zhe; position error correction is based on the coordinate prediction of the button to correct the possibility of a nearby button sex.

In actual use, due to different user equipment, input habits, etc., different input errors may occur. Therefore, whether it is the above-mentioned button error correction or position error correction, the factors considered in error correction are limited, and the accuracy of the error correction results. There is still room for improvement.

SUMMARY OF THE INVENTION

Embodiments of the present invention provide an input method error correction method and device, so as to improve the error correction capability and accuracy of the input method and improve the quality of user input.

For this reason, the present invention provides the following technical solutions:

An input method error correction method, the method comprises:

During the user input process, determine the grip type corresponding to the current button;

Obtain an error correction mapping table corresponding to the grip type;

Perform error correction processing according to the error correction mapping relationship table to obtain one or more input strings after error correction.

Preferably, the determining the grip type corresponding to the current key includes:

Acquire the recognition feature of the grip posture when the current button is triggered; the recognition feature of the grip posture includes any one or more of the following: position information, time information, motion information, device orientation, and device type;

The location information includes any one or more of the following: key coordinates, the distance between two adjacent keys, and the area where the key is located on the keyboard;

The time information includes: the time difference between two adjacent keys;

The motion information includes: acceleration, rotational angular velocity;

The grip posture recognition feature is input into a pre-built grip posture recognition model, and the grip posture type corresponding to the current key is determined according to the output of the grip posture recognition model.

Preferably, the method further comprises: constructing a grip gesture recognition model in the following manner:

Determine the topology of the grip gesture recognition model;

Collect training data, where the training data includes: grip posture recognition features under different grip postures during user input;

The parameters of the grip posture recognition model are obtained by training using the training data.

Preferably, the gripping posture types include: left-hand gripping the phone with left-hand input, right-hand gripping phone with right-hand input, and double-hand gripping phone with single thumb input.

Preferably, the method further includes:

The error correction mapping table corresponding to each grip type is constructed as follows:

Determine each button whose control distance is greater than the set value under the grip type, and use it as the button to be corrected;

Record and count the error correction information of the to-be-corrected keys, where the error-correction information includes: other keys replaced by the to-be-corrected keys and the number of times of replacement during user input;

An error correction mapping relationship table corresponding to the grip type is established according to the error correction information of the button to be corrected.

Preferably, the mapping relationship in the error correction mapping relationship table is a one-way mapping relationship.

Preferably, performing error correction processing according to the error correction mapping relationship table, and obtaining one or more input strings after error correction includes:

If there is a mapping relationship corresponding to the current key in the error correction mapping relationship table, use the character in the mapping relationship to replace the character corresponding to the current key in the input string to obtain one or more input after error correction string.

Preferably, performing error correction processing according to the error correction mapping relationship table, and obtaining one or more input strings after error correction includes:

If there is a mapping relationship corresponding to the current key in the error correction mapping relationship table, then use the character in the mapping relationship to replace the character corresponding to the current key in the input string to obtain the replaced input string;

Calculate the scores of the input string before the replacement character and the input string after the replacement character respectively;

The input string with higher score is obtained as the error-corrected input string.

Preferably, the method further includes:

Record the usage habits of each key in the user input process;

Before performing error correction processing according to the error correction mapping relationship table, judge whether it is necessary to perform error correction processing on the character that triggers the input of the current button according to the recorded usage habit of the current button;

If necessary, the step of performing error correction processing according to the error correction mapping relationship table is performed.

Preferably, the usage habit of the key includes: the number of times the key is pressed incorrectly and the total number of times the key is used;

The judging whether it is necessary to perform error correction processing on the character that triggers the input of the current key according to the recorded usage habits of the current key includes:

Calculate the error rate of the current button according to the recorded number of times the current button is pressed incorrectly and the total number of times the current button is used;

If the error rate is greater than the set error threshold, it is determined that it is necessary to perform error correction processing on the character that triggers the current key input;

Otherwise, it is determined that it is not necessary to perform error correction processing on the character that triggers the current key input.

An input method error correction device, the device comprises:

The grip posture determination module is used to determine the grip posture type corresponding to the current button during the user input process;

a mapping relationship determination module, configured to obtain an error correction mapping relationship table corresponding to the grip posture type;

An error correction module, configured to perform error correction processing according to the error correction mapping relationship table to obtain one or more input strings after error correction.

Preferably, the grip posture determination module includes:

The feature acquisition unit is used to acquire the grip gesture recognition feature when the current button is triggered during the user input process; the grip gesture recognition feature includes any one or more of the following: position information, time information, motion information, device orientation, Equipment type;

The location information includes any one or more of the following: key coordinates, the distance between two adjacent keys, and the area where the key is located on the keyboard;

The time information includes: the time difference between two adjacent keys;

The motion information includes: acceleration and rotational angular velocity.

The grip posture detection unit is configured to input the grip posture recognition feature into a pre-built grip posture recognition model, and determine the grip posture type corresponding to the current key according to the output of the grip posture recognition model.

Preferably, the device further comprises:

A model building module for building the grip posture recognition model; the model building module includes:

a topology determination unit, used to determine the topology of the grip posture recognition model;

a data collection unit, configured to collect training data, the training data including: grip posture recognition features under different grip postures during user input;

The parameter training unit is used for obtaining the parameters of the grip posture recognition model by training the training data.

Preferably, the gripping posture types include: left-hand gripping the phone with left-hand input, right-hand gripping phone with right-hand input, and double-hand gripping phone with single thumb input.

Preferably, the device further comprises:

A relationship table establishment module is used to pre-build an error correction mapping relationship table corresponding to each grip type; the relationship table establishment module includes:

A button determination unit to be corrected, configured to determine each button whose control distance is greater than the set value under the grip type, and use it as the button to be corrected;

A statistical unit, configured to record and count the error correction information of the to-be-corrected keys, where the error-correction information includes: other keys replaced by the to-be-corrected keys and the number of times of replacement during user input;

A relationship establishing unit, configured to establish an error correction mapping relationship table corresponding to the grip posture type according to the error correction information of the button to be corrected.

Preferably, the mapping relationship in the error correction mapping relationship table is a one-way mapping relationship.

Preferably, the error correction module is specifically configured to, when there is a mapping relationship corresponding to the current key in the error correction mapping relationship table, use the characters in the mapping relationship to replace the corresponding character of the current key in the input string character to get one or more input strings after error correction.

Preferably, the error correction module includes:

A character replacement unit is used to replace the character corresponding to the current key in the input string with the character in the mapping relationship when there is a mapping relationship corresponding to the current key in the error correction mapping relationship table, and obtain the replaced input string;

a score calculation unit, used to calculate the scores of the input string before the replacement character and the input string after the replacement respectively;

The output unit is used to obtain an input string with a higher score as an error-corrected input string.

Preferably, the device further comprises:

The recording module is used to record the usage habits of each key in the user input process;

The error correction judgment module is used to judge whether the character that triggers the input of the current button needs to be processed according to the usage habit of the current button recorded by the recording module before the error correction module performs error correction processing according to the error correction mapping relationship table. Error correction processing; if necessary, trigger the error correction module to perform error correction processing according to the error correction mapping relationship table.

Preferably, the usage habit of the key includes: the number of times the key is pressed incorrectly and the total number of times the key is used;

The error correction judgment module is specifically used to calculate the error rate of the current button according to the number of times that the current button is wrongly pressed and the total number of times the current button is used; if the error rate is greater than the set error If the threshold is set, it is determined that error correction processing needs to be performed on the character that triggers the input of the current key; otherwise, it is determined that the character that triggers the input of the current key does not need to be subjected to error correction.

A computer device, comprising: one or more processors, memory;

The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the methods described above.

A readable storage medium having stored thereon instructions that are executed to implement the aforementioned method.

In the input method error correction method and device provided by the embodiments of the present invention, in the process of user input, the holding posture type is determined by detecting the user's holding posture, and then the error correction is performed according to the error correction mapping relationship table corresponding to the holding posture type. After processing, one or more input strings after error correction are obtained, which effectively improves the error correction capability and accuracy of the input method, and improves the input quality of the user.

Description of drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the accompanying drawings required in the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only described in the present invention. For some of the embodiments, those of ordinary skill in the art can also obtain other drawings according to these drawings.

Fig. 1 is a flowchart of constructing a grip posture recognition model in an input method error correction method according to an embodiment of the present invention;

2 is a flowchart of an input method error correction method according to an embodiment of the present invention;

3 is an example of a keyboard layout in an input method error correction method according to an embodiment of the present invention;

4 is another flowchart of an input method error correction method according to an embodiment of the present invention;

5 is a schematic structural diagram of an input method error correction device according to an embodiment of the present invention;

6 is a schematic structural diagram of a model building module in an input method error correction device according to an embodiment of the present invention;

7 is another schematic structural diagram of an input method error correction device according to an embodiment of the present invention;

8 is a block diagram of an apparatus for an input method error correction method according to an exemplary embodiment;

FIG. 9 is a schematic structural diagram of a server in an embodiment of the present invention.

Detailed ways

In order to enable those skilled in the art to better understand the solutions of the embodiments of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

Considering that when the user performs input, the gesture of the user holding the input device will also affect the user's operation of the keys to a large extent, especially in the case of single-handed input, which may easily cause key input errors. In response to this situation, the embodiments of the present invention provide an input method error correction method and device. During the user input process, the grip posture type is determined by detecting the user's grip posture, and then the error correction corresponding to the grip posture type is performed. Error correction processing is performed on the mapping relationship table to obtain one or more input strings after error correction.

When performing grip posture detection, a pre-built grip posture recognition model can be used to input the acquired grip posture recognition features when each key is triggered into the grip posture recognition model, and the grip posture type corresponding to the key is determined according to the output of the model. Of course, the specific grip type may also be determined based on other information, such as feature sensing, etc., which is not limited in this embodiment of the present invention.

As shown in Figure 1, it is a flowchart of building a grip posture recognition model in an embodiment of the present invention, including the following steps:

Step 101: Determine the topology of the grip posture recognition model.

The topological structure of the grip posture recognition model can be adopted but is not limited to any of the following: DNN (Deep Neural Networks, deep neural network), RNN (Recurrent Neural Network, cyclic neural network), LSTM (Long Short-Term Memory, long and short term memory) memory network).

In practical applications, the grip posture recognition model may be a classification model or a regression model, which is not limited in this embodiment of the present invention.

The input of the grip gesture recognition model is the grip gesture recognition feature, and the output of the grip gesture recognition model will be different depending on whether the model adopts a classification model or a regression model. Specifically, if a classification model is used, the output of the model is the judgment result of the grip posture type. If the regression model is used, the output of the model is the probability of each grip type. The larger the probability value, the higher the possibility of the grip type. Therefore, according to the probability and the set probability threshold, the corresponding current The grip type of the button. Specifically, if the probability of a certain grip type is greater than the set probability threshold, the grip type of the current button is the grip type.

Step 102 , collecting training data, where the training data includes: grip posture recognition features under different grip postures in the user input process.

The grip gesture recognition features include, but are not limited to, any one or more of the following: location information, time information, motion information, device orientation, and device type. in:

The location information includes any one or more of the following: key coordinates, the distance between two adjacent keys, and the area where the key is located on the keyboard. The distance between the two adjacent keys refers to the distance between the current key and the previous key, and the distance can be calculated according to the coordinates of the two keys; the area where the key is located on the keyboard refers to the current key Whether it is on the left or right side of the keyboard. Although the layout of keys on the keyboard and the positions on the screen may be different for different input devices, these information can be obtained through the application program interface provided by the input device.

The time information includes: the time difference between two adjacent keys. Specifically, the time difference between the current key press and the previous key press can be obtained by recording the time triggered by each key press and subtracting the two.

The motion information includes: acceleration and rotational angular velocity. This information can also be obtained through the application program interface provided by the input device. Usually, the input device will have its own accelerometer and gyroscope to sense these two parameters respectively.

It should be noted that the training data may come from the input of multiple different users using various gripping postures, and the amount of training data should be sufficient, because the more training samples, the more accurate the model parameters obtained by training. .

Step 103, using the training data to train to obtain the parameters of the grip posture recognition model.

The training method of the model parameters can adopt the existing technology, and the principle is as follows:

1) Initialize the weights, for example, a normal distribution can be used to randomize the initial weights;

2) Select a sample set, that is, a sample (Ai, Bi) of the training data, wherein Ai is the training data, that is, the grip posture recognition feature; Bi is a label, that is, the gripping posture corresponding to Ai;

3) Input Ai into the network, and calculate the output Y of the network;

4) Calculate the error between the predicted value and the actual value: D=Bi-Y;

5) Adjust the weight matrix W according to the error D;

6) Repeat the above process for each sample until the error does not exceed the specified range for the entire sample set.

Through the above process, the weight matrix W of the network is finally obtained, that is, the parameters of the grip posture recognition model.

It should be noted that, in the process of inputting by the user using the input method using the error correction scheme of the present invention, the data required for model training can also be collected, and the parameters of the grip posture recognition model can be updated by using these data, such as When the collected data reaches a certain amount, retrain the grip recognition model together with all or part of the original training data to obtain the updated model parameters; or use the weight of the pre-trained model as the initialization weight , and then adjust the weights for the new dataset, training faster than starting with random initialization, and resulting in better model parameter accuracy.

In the input method error correction method provided by the embodiment of the present invention, during the user input process, the grip type corresponding to each button is detected, and then the error correction mapping table corresponding to the grip type is used to correct the input wrong button. .

As shown in Figure 2, it is a flowchart of an input method error correction method according to an embodiment of the present invention, including the following steps:

Step 201 , during the user input process, determine the type of grip posture corresponding to the current key.

For example, the current grip type can be determined based on the above-mentioned grip gesture recognition model. Specifically, first, the grip posture recognition feature when the current button is triggered is obtained, then the grip posture recognition feature is input into a pre-built grip posture recognition model, and the grip corresponding to the current button is determined according to the output of the grip posture recognition model. pose type.

The grip gesture recognition feature is the same as the grip gesture recognition feature used in the construction of the previous grip gesture recognition model, and may specifically include, but is not limited to, any one or more of the following: position information, time information, motion information, device orientation, equipment type. The location information includes any one or more of the following: key coordinates, the distance between two adjacent keys, and the area where the key is located on the keyboard; the time information includes: the time difference between two adjacent keys; the The motion information includes: acceleration, rotational angular velocity.

The specific meanings and how to obtain the above-mentioned grip posture recognition features have been described in detail above, and will not be repeated here.

As mentioned earlier, the input of the grip gesture recognition model is the grip gesture recognition feature, and the output of the grip gesture recognition model will be different depending on whether the model adopts a classification model or a regression model. Specifically, if a classification model is used, the output of the model is the judgment result of the grip posture type. If the regression model is used, the output of the model is the probability of each grip type. The larger the probability value, the higher the possibility of the grip type. Therefore, according to the probability and the set probability threshold, the corresponding current The grip type of the button. Specifically, if the probability of a certain grip type is greater than the set probability threshold, the grip type of the current button is the grip type.

Considering that when the user is inputting, the phenomenon of key input errors caused by the holding posture is mainly the case of single-handed input. Therefore, in practical applications, the holding posture types may include: left-hand holding the phone for left-hand input, right-hand holding the phone for right-hand input. , Hold the machine with both hands and input with a single thumb. Of course, more detailed distinctions can also be made. For example, in addition to the above-mentioned types of holding postures, any one or more of the following types may be further included: two-hand input with both hands, left-hand with right-hand input, right-hand with left-hand input , device fixed one-hand input, etc.

Step 202: Acquire an error correction mapping relationship table corresponding to the grip posture type.

The error correction mapping relationship table may be constructed in advance according to the distribution and positional relationship of keys in the keyboard, and statistical information of related keys, and may be constructed automatically or manually, which is not limited in this embodiment of the present invention. The error correction mapping relationship table reflects the mapping relationship of the keys that are easy to enter wrongly under the condition of the corresponding holding posture.

Specifically, a construction method of the error correction mapping relationship table corresponding to each grip posture type is as follows:

Determine each button whose control distance is greater than the set value under the grip type, and use it as the button to be corrected;

Record and count the error correction information of the to-be-corrected keys, where the error-correction information includes: other keys replaced by the to-be-corrected keys and the number of times of replacement during user input;

An error correction mapping relationship table corresponding to the grip type is established according to the error correction information of the button to be corrected. For example, for the keyboard layout shown in Figure 3, corresponding to the case where the grip type is the right hand holding the phone and the right hand input, the keys in the diagonally diagonal keyboard area of the right hand can be used as the basis, because the control distance of the keys in this area is relatively long. For example, keys q, w, e, a, s, d, z, etc., are used as the keys to be corrected, the error correction information of these keys is recorded and counted, and a mapping relationship is established according to the error correction information.

It should be noted that the mapping relationship in the error correction mapping relationship table is different from that of ordinary error correction. It is a one-way mapping relationship, for example, the mapping relationship is s->q, and in practical applications, according to the keyboard Depending on the layout, the mapping relationship may be a one-to-one mapping relationship or a one-to-many mapping relationship.

Step 203: Perform error correction processing according to the error correction mapping relationship table to obtain one or more input strings after error correction.

Specifically, when performing error correction processing, it is possible to first check whether there is a mapping relationship corresponding to the current key in the error correction mapping relationship table, and if so, use the characters in the mapping relationship to replace the current key in the input string. The character corresponding to the key. Since the mapping relationship may be one-to-one or one-to-many, there may also be one or more input strings after error correction. Of course, if there is no mapping relationship corresponding to the current key in the error correction mapping relationship table, the above-mentioned character replacement is not required, and the next key input is continued to be detected.

For example, when the user holds the phone with the right hand and inputs with the right hand, and uses the Pinyin input method to input "Han Qin", the string to be input is hanqin. became hansin. Since it is detected in advance that the user is inputting with the right hand holding the phone, the error correction mapping relationship corresponding to the right hand input is established. Therefore, according to the mapping relationship s->q, the input character "s" is replaced with "q", The error-corrected input string "hanqin" is obtained, and the candidate word "hanqin" corresponding to the input string "hanqin" is output in the candidate column.

It should be noted that the user's input process is continuous, and accordingly, the above error correction process needs to be performed for each key input.

Further, in order to avoid erroneous replacement, in another embodiment of the method of the present invention, when performing error correction processing, the character corresponding to the current key in the input string can be replaced first according to the aforementioned error correction processing method, so as to obtain: The input string after the replacement; then calculate the score of the input string before the replacement character and the input string after the replacement character respectively, and the score can specifically be the sorting score in the input method; Obtain the input string with a higher score as the error correction the input string.

In the input method error correction method provided by the embodiment of the present invention, during the user input process, the grip type corresponding to the current button is detected, and then the error correction mapping table corresponding to the grip type is used to correct the input wrong button. , thereby effectively improving the error correction ability and accuracy of the input method, especially for the key input errors caused by the user's single-handed holding the machine input, a good error correction effect can be obtained, and the user input quality is greatly improved.

As shown in Figure 4, it is another flowchart of the input method error correction method according to the embodiment of the present invention, including the following steps:

In step 401, the user starts to input.

Step 402: Determine the type of grip posture corresponding to the current key.

The gripping posture types may include: left-hand gripping the phone for left-hand input, right-hand gripping the phone for right-hand input, and two-hand gripping the phone for single-thumb input. Of course, more detailed distinctions can be made. For example, in addition to the above-mentioned types of gripping postures, any one or more of the following types can be further included: two-hand input with two hands, two-hand input with left-hand grip, right-hand input with right-hand grip and left-hand input , device fixed one-hand input, etc.

Step 403: Obtain an error correction mapping relationship table corresponding to the grip posture type.

Step 404 , check whether there is a mapping relationship corresponding to the current key in the error correction mapping relationship table; if yes, go to step 405 ; otherwise, return to step 402 .

The mapping relationship may be one-to-one or one-to-many. In step 404, in the one-to-many case, as long as it is detected that there is a mapping relationship corresponding to the current key, step 405 is executed.

Step 405 , according to the recorded usage habits of the current key, determine whether it is necessary to perform error correction processing on the character that triggers the input of the current key; if yes, go to step 406 ; otherwise, return to step 402 .

The key usage habit may include: the number of wrong key presses and the total number of times the key is used. The usage habits of each key can be recorded and counted during the user's input process.

When judging whether it is necessary to perform error correction processing on the character that triggers the input of the current key, the error rate of the current key can be calculated according to the recorded number of times the wrong key is pressed and the total number of times the current key is used; If the error rate is greater than the set error threshold, it is determined that error correction processing needs to be performed on the character that triggers the input of the current key; otherwise, it is determined that the character that triggers the input of the current key does not need to be corrected.

Step 406: Perform error correction processing according to the error correction mapping relationship table to obtain one or more input strings after error correction.

The error correction processing mainly uses the characters in the mapping relationship to replace the characters corresponding to the current keys in the input string. Similarly, in order to avoid wrong replacement, when performing error correction processing, the scores of the input string before the replacement character and the input string after the replacement character can also be calculated respectively, and the score can be the sorting score in the input method; obtain the score; The higher input string is used as the error-corrected input string. Of course, if the mapping relationship is a one-to-many relationship, multiple input strings after error correction can also be obtained.

Compared with the embodiment shown in FIG. 2, the input method error correction method of this embodiment records the usage habit of each key in the user input process, and comprehensively considers the historical operation of each key during user input when performing error correction processing. It further improves the accuracy of error correction and improves the user's input experience.

It should be noted that the input method error correction method provided by the embodiment of the present invention can be applied to various input methods, and is not only applicable to Chinese input realized by various input methods, but also applicable to English and other languages If the input is in English, the final output candidate word is an English word.

Correspondingly, an embodiment of the present invention further provides an input method error correction apparatus, which may be integrated in user equipment, and the user equipment may be a notebook, a computer, a PAD, a mobile phone, or the like. When a user performs an input operation, the user needs to press or touch a corresponding key in the keyboard of the user equipment, and the keyboard may be a physical keyboard or a virtual keyboard on the touch screen of the user equipment.

As shown in FIG. 5 , it is a schematic structural diagram of an input method error correction apparatus according to an embodiment of the present invention.

In this embodiment, the apparatus includes:

The grip posture determination module 501 is used for determining the grip posture type corresponding to the current key during the user input process;

a mapping relationship determining module 502, configured to obtain an error correction mapping relationship table corresponding to the grip posture type;

The error correction module 503 is configured to perform error correction processing according to the error correction mapping relationship table to obtain one or more input strings after error correction.

The above-mentioned grip posture determination module 501 can specifically determine the type of the grip posture corresponding to the current key in a model-based manner. Of course, it can also determine the specific type of grip posture based on other information, such as feature sensing, etc., which is not done in this embodiment of the present invention. limited.

A specific implementation structure of the grip posture determination module 501 may include: a feature acquisition unit and a grip posture detection unit (not shown). Wherein, the feature acquisition unit is used to acquire the grip posture recognition feature when the current button is triggered during the user input process; the grip posture detection unit is used to input the grip posture recognition feature into a pre-built grip posture recognition model, According to the output of the grip gesture recognition model, the grip gesture type corresponding to the current key is determined.

The grip gesture recognition feature is the same as the grip gesture recognition feature used in the construction of the previous grip gesture recognition model, and may specifically include but not limited to any one or more of the following: position information, time information, motion information, device orientation, equipment type. The location information includes any one or more of the following: key coordinates, the distance between two adjacent keys, and the area where the key is located on the keyboard; the time information includes: the time difference between two adjacent keys; the The motion information includes: acceleration, rotational angular velocity.

The specific meanings and how to obtain the above-mentioned grip posture recognition features have been described in detail above, and will not be repeated here.

The grip gesture detection unit inputs the grip gesture recognition feature into a pre-built grip gesture recognition model, and determines the grip gesture type corresponding to the current key according to the output of the grip gesture recognition model. Specifically, the grip posture recognition model may be a classification model or a regression model. If a regression model is used, the output of the grip posture recognition model is the probability of each grip posture type; the grip posture detection unit needs to further The probability of the gesture type is compared with the set probability threshold. If the probability of a certain grip type is greater than the set probability threshold, the grip type of the current button is the grip type.

In practical applications, the gripping posture types may include: left-hand gripping the phone for left-hand input, right-hand gripping the phone for right-hand input, and two-hand gripping the phone for single-thumb input. Of course, more detailed distinctions can be made. For example, in addition to the above-mentioned types of gripping postures, any one or more of the following types can be further included: two-hand input with two hands, two-hand input with left-hand grip, right-hand input with right-hand grip and left-hand input , device fixed one-hand input, etc.

The error correction mapping relationship table may be constructed manually or by the system in advance according to the distribution and positional relationship of keys in the keyboard, and the error correction mapping relationship table includes the mapping relationship of keys that are easy to enter wrongly under the condition of the corresponding holding posture. The mapping relationship may be a one-to-one relationship or a one-to-many relationship, and the mapping relationship is unidirectional. For example, a relationship table establishment module (not shown) can be set up, and the error correction mapping relationship table corresponding to each grip posture type can be pre-built. The relationship table establishment module can be integrated into the input method error correction device of the present invention, and can be independent of The input method error correction device of the present invention is not limited thereto.

A specific structure of the relationship table establishment module may include the following units:

A button determination unit to be corrected, configured to determine each button whose control distance is greater than the set value under the grip type, and use it as the button to be corrected;

A statistical unit, configured to record and count the error correction information of the to-be-corrected keys, where the error-correction information includes: other keys replaced by the to-be-corrected keys and the number of times of replacement during user input;

A relationship establishing unit, configured to establish an error correction mapping relationship table corresponding to the grip posture type according to the error correction information of the button to be corrected.

The error correction processing of the current button by the error correction module 503 is mainly to check whether there is a mapping relationship corresponding to the current button in the error correction mapping relationship table, and if so, use the characters in the mapping relationship to replace all the characters in the input string. The character corresponding to the current key is described, and one or more input strings after error correction are obtained. Of course, if there is no mapping relationship corresponding to the current key in the error correction mapping relationship table, the above-mentioned character replacement is not required.

It should be noted that the user's input process is continuous, and accordingly, the above error correction process needs to be performed for each key input.

In order to avoid mistaken replacement, the error correction module 503 can further calculate the scores of the input string before the replacement character and the input string after the replacement respectively, and the score can specifically be the ranking score in the input method, and input the input string with higher score. string as the error-corrected input string. Correspondingly, a specific structure of the error correction module 503 may include the following units:

A character replacement unit is used to replace the character corresponding to the current key in the input string with the character in the mapping relationship when there is a mapping relationship corresponding to the current key in the error correction mapping relationship table, and obtain the replaced input string;

a score calculation unit, used to calculate the scores of the input string before the replacement character and the input string after the replacement respectively;

The output unit is used to obtain an input string with a higher score as an error-corrected input string.

In the input method error correction device provided by the embodiment of the present invention, in the process of user input, the grip type corresponding to each key is detected, and then the error correction mapping table corresponding to the grip type is used to correct the input wrong key. , thereby effectively improving the error correction ability and accuracy of the input method, especially for the key input errors caused by the user's single-handed holding the machine input, a good error correction effect can be obtained, and the user input quality is greatly improved.

The grip posture recognition model can be constructed by a model building module using a large amount of training data collected. The model building module can be integrated into the input method error correction device of the present invention, and can be independent of the input method error correction device of the present invention. limited.

As shown in FIG. 6 , it is a schematic structural diagram of a model building module in an input method error correction device according to an embodiment of the present invention.

The model building block includes the following units:

The topology determination unit 61 is configured to determine the topology of the grip gesture recognition model; the topology may be, but not limited to, any of the following: DNN, RNN, LSTM, and the like.

The data collection unit 62 is configured to collect training data, where the training data includes: grip posture recognition features under different grip postures in the user input process. The training data may come from situations where multiple different users use various gripping postures for input, and the gripping posture identification features have been described in detail above, and will not be repeated here.

The parameter training unit 63 is configured to use the training data to train to obtain the parameters of the grip posture recognition model. The training method of the model parameters may adopt the prior art, which will not be repeated here.

It should be noted that, in practical applications, the above-mentioned data collection unit 63 may also collect data required for model training during each input process by the user, and use these data to update the parameters of the grip posture recognition model, such as collecting When the data reaches a certain amount, the parameter training unit 64 is triggered to retrain the grip posture recognition model together with all or part of the original training data to obtain updated model parameters; or trigger the parameter training The unit 64 uses the weights of the pre-trained model as the initialization weights, and then adjusts the weights for the new data set. Compared with starting from random initialization, the training can be faster and the accuracy of the obtained model parameters is better.

As shown in FIG. 7 , it is another schematic structural diagram of an input method error correction apparatus according to an embodiment of the present invention.

The difference from the embodiment shown in FIG. 5 is that in this embodiment, the apparatus not only includes the aforementioned holding posture determination module 501, mapping relationship determination module 502, and error correction module 503, but also further includes: a recording module 504 and an error correction module. Judgment module 505 . in:

The recording module 504 is used to record the usage habits of each key in the user input process;

The error correction judgment module 505 is configured to judge whether it is necessary to trigger the current button according to the usage habit of the current button recorded by the recording module 504 before the error correction module 503 performs error correction processing according to the error correction mapping relationship table. The inputted characters are subjected to error correction processing; if necessary, the error correction module 503 is triggered to perform error correction processing according to the error correction mapping relationship table.

Wherein, the usage habit of the key may include: the number of times the key is pressed incorrectly and the total number of times the key is used.

Correspondingly, the error correction judgment module 505 can specifically calculate the error rate of the current button according to the number of times the current button is pressed incorrectly and the total number of times the current button is used; if the error rate is greater than the set value. If the error threshold is determined, it is determined that error correction processing needs to be performed on the character that triggers the input of the current key; otherwise, it is determined that the character that triggers the input of the current key does not need to be subjected to error correction.

Regarding the apparatuses in the foregoing embodiments, reference may be made to the descriptions in the embodiments of the method for specific manners in which each module performs operations, which will not be described in detail here.

It can be seen that the input method error correction device of the embodiment of the present invention further improves the error correction by recording the usage habit of each button during the user input process, and comprehensively considering the historical operation habit of each button during the user input when performing error correction processing. accuracy, improving the user's input experience.

The input method error correction device provided in the embodiment of the present invention can be applied to various input methods, and is not only applicable to Chinese input realized by various input methods, but also applicable to input in English and other languages, such as English enter.

It should be noted that the input method error correction method and device provided by the embodiments of the present invention can be used independently, or can be combined with other existing error correction schemes as an optimization scheme of other existing error correction schemes, that is, The error correction based on grip posture detection is performed first, and then the error correction of other schemes is performed, so that the error correction capability and accuracy of the input method can be further improved, and the input quality of the user can be improved.

The method and apparatus of the embodiments of the present invention may be applied to a physical keyboard or a virtual keyboard on a screen of a user equipment, and the user equipment may be a computer, a notebook, a mobile phone, a PAD, and the like.

FIG. 8 is a block diagram of an apparatus 800 for an input method error correction method according to an exemplary embodiment. For example, apparatus 800 may be a mobile phone, computer, digital broadcast terminal, messaging device, game console, tablet device, medical device, fitness device, personal digital assistant, and the like.

8, the apparatus 800 may include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814, and communication component 816.

The processing component 802 generally controls the overall operation of the device 800, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing element 802 may include one or more processors 820 to execute instructions to perform all or part of the steps of the methods described above. Additionally, processing component 802 may include one or more modules that facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802.

Memory 804 is configured to store various types of data to support operation at device 800 . Examples of such data include instructions for any application or method operating on device 800, contact data, phonebook data, messages, pictures, videos, and the like. Memory 804 may be implemented by any type of volatile or nonvolatile storage device or combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic or Optical Disk.

Power component 806 provides power to various components of device 800 . Power components 806 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to device 800 .

Multimedia component 808 includes a screen that provides an output interface between the device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, swipe, and gestures on the touch panel. The touch sensor may not only sense the boundaries of a touch or swipe action, but also detect the duration and pressure associated with the touch or swipe action. In some embodiments, the multimedia component 808 includes a front-facing camera and/or a rear-facing camera. When the device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each of the front and rear cameras can be a fixed optical lens system or have focal length and optical zoom capability.

Audio component 810 is configured to output and/or input audio signals. For example, audio component 810 includes a microphone (MIC) that is configured to receive external audio signals when device 800 is in operating modes, such as call mode, recording mode, and voice recognition mode. The received audio signal may be further stored in memory 804 or transmitted via communication component 816 . In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

The I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module, which may be a keyboard, a click wheel, a button, or the like. These buttons may include, but are not limited to: home button, volume buttons, start button, and lock button.

Sensor assembly 814 includes one or more sensors for providing status assessment of various aspects of device 800 . For example, the sensor assembly 814 can detect the open/closed state of the device 800, the relative positioning of components, such as the display and keypad of the device 800, and the sensor assembly 814 can also detect a change in the position of the device 800 or a component of the device 800 , the presence or absence of user contact with the device 800 , the orientation or acceleration/deceleration of the device 800 and the temperature change of the device 800 . Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact. Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

Communication component 816 is configured to facilitate wired or wireless communication between apparatus 800 and other devices. Device 800 may access wireless networks based on communication standards, such as WiFi, 2G or 3G, or a combination thereof. In one exemplary embodiment, the communication component 816 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

In an exemplary embodiment, apparatus 800 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable A gate array (FPGA), controller, microcontroller, microprocessor or other electronic component implementation is used to perform the above method.

In an exemplary embodiment, there is also provided a non-transitory computer-readable storage medium including instructions, such as a memory 804 including instructions, executable by the processor 820 of the apparatus 800 to perform the method described above. For example, the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like.

The present invention also provides a non-transitory computer-readable storage medium, when the instructions in the storage medium are executed by the processor of the mobile terminal, the mobile terminal can perform some or all of the steps in the above method embodiments, so as to Implement error correction for corresponding input errors.

FIG. 9 is a schematic structural diagram of a server in an embodiment of the present invention. The server 1900 may vary greatly due to different configurations or performances, and may include one or more central processing units (CPUs) 1922 (eg, one or more processors) and a memory 1932, one or more A storage medium 1930 (eg, one or more mass storage devices) that stores applications 1942 or data 1944 above. Among them, the memory 1932 and the storage medium 1930 may be short-term storage or persistent storage. The program stored in the storage medium 1930 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Furthermore, the central processing unit 1922 may be configured to communicate with the storage medium 1930 to execute a series of instruction operations in the storage medium 1930 on the server 1900 .

Server 1900 may also include one or more power supplies 1926, one or more wired or wireless network interfaces 1950, one or more input/output interfaces 1958, one or more keyboards 1956, and/or, one or more operating systems 1941 , such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM and many more.

Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. The present invention is intended to cover any modifications, uses or adaptations of the present invention that follow the general principles of the invention and include common knowledge or common technical means in the art not disclosed by this disclosure . The specification and examples are to be regarded as exemplary only, with the true scope and spirit of the invention being indicated by the following claims.

It should be understood that the present invention is not limited to the precise structures described above and illustrated in the accompanying drawings, and that various modifications and changes may be made without departing from its scope. The scope of the present invention is limited only by the appended claims

The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within the range.

Claims (10)

1. An input method error correction method, characterized in that the method comprises:

determining a holding posture type corresponding to a current key in a user input process;

acquiring an error correction mapping relation table corresponding to the holding posture type;

and carrying out error correction processing according to the error correction mapping relation table to obtain one or more input strings after error correction.

2. The method of claim 1, wherein the determining the grip type corresponding to the current key comprises:

acquiring a holding posture identification characteristic when a current key is triggered; the gripping posture identification features comprise any one or more of the following: location information, time information, motion information, device direction, device type;

the location information includes any one or more of: the key coordinates, the distance between two adjacent keys and the area of the keys on the keyboard;

the time information includes: time difference of two adjacent keys;

the motion information includes: acceleration, rotational angular velocity;

and inputting the gripping posture identification characteristics into a pre-constructed gripping posture identification model, and determining a gripping posture type corresponding to the current key according to the output of the gripping posture identification model.

3. The method of claim 1, further comprising:

constructing an error correction mapping relation table corresponding to each holding posture type according to the following modes:

determining each key of which the operating distance is greater than a set value under the holding posture type, and taking the key as a key to be corrected;

recording and counting error correction information of the key to be corrected, wherein the error correction information comprises: other keys and replacement times which are replaced by the keys to be corrected in the input process of the user;

and establishing an error correction mapping relation table corresponding to the holding posture type according to the error correction information of the key to be corrected.

4. The method according to claim 1, wherein said performing error correction processing according to the error correction mapping table to obtain one or more corrected input strings comprises:

and if the error correction mapping relation table has the mapping relation corresponding to the current key, replacing the character corresponding to the current key in the input string with the character in the mapping relation to obtain one or more corrected input strings.

5. The method according to claim 1, wherein said performing error correction processing according to the error correction mapping table to obtain one or more corrected input strings comprises:

if the mapping relation corresponding to the current key exists in the error correction mapping relation table, replacing the character corresponding to the current key in the input string by using the character in the mapping relation to obtain a replaced input string;

respectively calculating scores of the input string before the character is replaced and the input string after the character is replaced;

and acquiring the input string with higher score as the input string after error correction.

6. An input method error correction apparatus, characterized in that the apparatus comprises:

the holding posture determining module is used for determining the holding posture type corresponding to the current key in the input process of the user;

the mapping relation determining module is used for acquiring an error correction mapping relation table corresponding to the holding posture type;

and the error correction module is used for carrying out error correction processing according to the error correction mapping relation table to obtain one or more input strings after error correction.

7. The apparatus of claim 6, wherein the grip posture determination module comprises:

the characteristic acquisition unit is used for acquiring holding posture identification characteristics when the current key is triggered in the input process of a user; the gripping posture identification features comprise any one or more of the following: location information, time information, motion information, device direction, device type;

the location information includes any one or more of: the key coordinates, the distance between two adjacent keys and the area of the keys on the keyboard;

the time information includes: time difference of two adjacent keys;

the motion information includes: acceleration, rotational angular velocity.

And the holding posture detection unit is used for inputting the holding posture identification characteristics into a pre-constructed holding posture identification model and determining the holding posture type corresponding to the current key according to the output of the holding posture identification model.

8. The apparatus of claim 6, further comprising:

the relation table establishing module is used for establishing an error correction mapping relation table corresponding to each holding posture type in advance; the relationship table establishing module comprises:

the key to be corrected determining unit is used for determining each key of which the operating distance is greater than a set value under the holding posture type and taking the key as the key to be corrected;

the statistical unit is used for recording and counting the error correction information of the key to be corrected, and the error correction information comprises: other keys and replacement times which are replaced by the keys to be corrected in the input process of the user;

and the relationship establishing unit is used for establishing an error correction mapping relationship table corresponding to the holding posture type according to the error correction information of the key to be corrected.

9. A computer device, comprising: one or more processors, memory;

the memory is for storing computer-executable instructions, and the processor is for executing the computer-executable instructions to implement the method of any one of claims 1 to 5.

10. A readable storage medium having stored thereon instructions that are executed to implement the method of any one of claims 1 to 5.

Images