METHOD AND APPARATUS FOR IMPROVING RECOGNITION ACCURACY FOR THE HANDWRITTEN INPUT OF ALPHANUMERIC CHARACTERS AND GESTURES

Abstract

A method for automatically selecting one of a plurality of recognition algorithms for a handwritten input of alphanumeric characters and/or gestures into a selected input field displayed on a screen using a touch-sensitive input apparatus comprises carrying out optical character recognition in a region of the screen which comprises at least the input field and the immediate environment of the input field, or carrying out voice recognition for a voice instruction acoustically output after the selected input field has been displayed. Terms describing field types are searched for in the result of the optical character recognition or the voice recognition, and a recognition algorithm which is adapted to a field type found in the result of the optical character recognition or the voice recognition is selected.

Claims

1. A method for automatically selecting one of a plurality of recognition algorithms or one of a plurality of parameter sets for a recognition algorithm for a handwritten input of alphanumeric characters and/or gestures into a selected input field displayed on a screen using a touch-sensitive input apparatus, comprising: carrying out optical character recognition in a region of the screen which includes at least the input field and the immediate environment of the input field, or carrying out voice recognition for a voice instruction acoustically output after the selected input field has been displayed, searching for terms in the result of the optical character recognition or the voice recognition, on the basis of which terms the field type of the input field can be determined, determining the field type on the basis of the terms found, and selecting a recognition algorithm which is adapted to a field type found in the result of the optical character recognition or the voice recognition or a parameter set for the recognition algorithm.

2. The method as claimed in claim 1, wherein the optical character recognition comprises: transmitting an image of what is displayed in at least one region of the screen which includes at least the input field and the immediate environment of the input field to an apparatus or to a computer program for optical character recognition, and receiving the result of the optical character recognition.

3. The method as claimed in claim 1, wherein the voice recognition comprises: recording the acoustically output voice instruction or receiving a signal representing the acoustic voice instruction, and receiving the result of the voice recognition.

4. The method as claimed in claim 3, wherein the signal representing the acoustic voice instruction is a digital or analog representation of electrical signals output via one or more loudspeakers or a control signal for a text-to-speech output unit.

5. The method as claimed in claim 1, wherein the optical character recognition or the voice recognition is carried out after one of a plurality of input fields has been selected on the screen.

6. An apparatus for automatically selecting one of a plurality of recognition algorithms for a handwritten input of alphanumeric characters and/or gestures into a selected input field displayed on a screen using a touch-sensitive input apparatus, comprising: first means which are set up to carry out optical character recognition in a region of the screen which includes at least the input field and the immediate environment of the input field, or are set up to carry out voice recognition for a voice instruction acoustically output after the selected input field has been displayed, second means which are set up to search for terms describing field types in the result of the optical character recognition or the voice recognition, and third means which are set up to select a recognition algorithm which is adapted to a field type found in the result of the optical character recognition or the voice recognition or a parameter set for the recognition algorithm.

7. The apparatus as claimed in claim 6, also comprising a fourth means which is set up to select a recognition algorithm or a parameter set for the recognition algorithm on the basis of possible characters or gestures to be input, wherein the possible characters or gestures to be input are determined from preceding inputs, and wherein the determination is carried out according to linguistic rules or by comparing words stored in a database.

8. The apparatus as claimed in claim 7, wherein one or more of the first, second, third and/or fourth means have one or more microprocessors and main memories and non-volatile memories communicatively connected to the one or more microprocessors, wherein respective non-volatile memories store computer program instructions which, when loaded into the respective main memory from the one or more microprocessors and executed, cause the performance of comprising: carrying out optical character recognition in a region of the screen which includes at least the input field and the immediate environment of the input field, or carrying out voice recognition for a voice instruction acoustically output after the selected input field has been displayed, searching for terms in the result of the optical character recognition or the voice recognition, on the basis of which terms the field type of the input field can be determined, determining the field type on the basis of the terms found, and selecting a recognition algorithm which is adapted to a field type found in the result of the optical character recognition or the voice recognition or a parameter set for the recognition algorithm, wherein the first, second and third means together perform all steps of the method.

9. The apparatus as claimed in claim 6, wherein the screen, the first, second, third and/or fourth means are arranged in a manner spatially separated from one another and are connected to one another by means of one or more communication networks.

10. A motor vehicle having an apparatus for automatically selecting one of a plurality of recognition algorithms for a handwritten input of alphanumeric characters and/or gestures into a selected input field displayed on a screen using a touch-sensitive input apparatus, comprising: first means which are set up to carry out optical character recognition in a region of the screen which includes at least the input field and the immediate environment of the input field, or are set up to carry out voice recognition for a voice instruction acoustically output after the selected input field has been displayed, second means which are set up to search for terms describing field types in the result of the optical character recognition or the voice recognition, and third means which are set up to select a recognition algorithm which is adapted to a field type found in the result of the optical character recognition or the voice recognition or a parameter set for the recognition algorithm.

11. The apparatus as claimed in claim 8, wherein the optical character recognition comprises: transmitting an image of what is displayed in at least one region of the screen which includes at least the input field and the immediate environment of the input field to an apparatus or to a computer program for optical character recognition, and receiving the result of the optical character recognition.

12. The apparatus as claimed in claim 8, wherein the voice recognition comprises: recording the acoustically output voice instruction or receiving a signal representing the acoustic voice instruction, and receiving the result of the voice recognition.

13. The method as claimed in claim 12, wherein the signal representing the acoustic voice instruction is a digital or analog representation of electrical signals output via one or more loudspeakers or a control signal for a text-to-speech output unit.

14. The method as claimed in claim 8, wherein the optical character recognition or the voice recognition is carried out after one of a plurality of input fields has been selected on the screen.

15. The apparatus as claimed in claim 7, wherein the screen, the first, second, third and/or fourth means are arranged in a manner spatially separated from one another and are connected to one another by means of one or more communication networks.

16. The apparatus as claimed in claim 8, wherein the screen, the first, second, third and/or fourth means are arranged in a manner spatially separated from one another and are connected to one another by means of one or more communication networks.

17. The apparatus as claimed in claim 10, also comprising a fourth means which is set up to select a recognition algorithm or a parameter set for the recognition algorithm on the basis of possible characters or gestures to be input, wherein the possible characters or gestures to be input are determined from preceding inputs, and wherein the determination is carried out according to linguistic rules or by comparing words stored in a database.

18. The apparatus as claimed in claim 17, wherein one or more of the first, second, third and/or fourth means have one or more microprocessors and main memories and non-volatile memories communicatively connected to the one or more microprocessors, wherein respective non-volatile memories store computer program instructions which, when loaded into the respective main memory from the one or more microprocessors and executed, cause the performance of steps comprising: carrying out optical character recognition in a region of the screen which includes at least the input field and the immediate environment of the input field, or carrying out voice recognition for a voice instruction acoustically output after the selected input field has been displayed, searching for terms in the result of the optical character recognition or the voice recognition, on the basis of which terms the field type of the input field can be determined, determining the field type on the basis of the terms found, and selecting a recognition algorithm which is adapted to a field type found in the result of the optical character recognition or the voice recognition or a parameter set for the recognition algorithm, wherein the first, second and third means together perform all steps of the method.

19. The apparatus as claimed in claim 10, wherein the screen, the first, second, third and/or fourth means are arranged in a manner spatially separated from one another and are connected to one another by means of one or more communication networks.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The invention is described below on the basis of the figures in the drawing. In the drawing:

[0037] FIG. 1 shows a schematic illustration of a dashboard in a motor vehicle;

[0038] FIG. 2 shows a schematic illustration of a screen for explaining the method according to the invention;

[0039] FIG. 3 shows a schematic flowchart of the method according to the invention, and

[0040] FIG. 4 shows a schematic illustration of a block diagram of an apparatus according to the invention.

DETAILED DESCRIPTION

[0041] In the figures, the same or similar elements are provided with the same reference symbols.

[0042] FIG. 1 shows a schematic illustration of a dashboard 100 in a motor vehicle. An instrument cluster 106 is arranged behind a steering wheel 104. A screen 110 and a touch-sensitive touchpad 116 are arranged in the region of a center console 108. A region in which a head-up display 114 displays information is also shown on a windshield 112. Like the screen 110 and the head-up display 114, the instrument cluster 106 may comprise a matrix display with graphics capability, that is to say a display which displays complex graphical structures by accordingly controlling individual pixels. That is to say, input fields for user inputs can be displayed on any of the display apparatuses. Irrespective of the display apparatus on which an input field is displayed, a user can make a handwritten input using the touch-sensitive touchpad 116 or an accordingly marked region on the screen 110. An optional holographic representation of visual contents and an optional connection to a coupled smartphone or to a cloud server via an Internet connection, for example, are not shown in the figure.

[0043] FIG. 2 shows the schematic illustration of the screen 110 from FIG. 1 for explaining the method according to the invention. The starting situation is that in which a user would like to input a telephone number into an input field 202 provided for this purpose. A plurality of options are available to the user for this purpose. The user can input the digits by touching the corresponding virtual number keys of the numeric keypad 204 displayed on the screen 110. It should be noted that the screen 110 is a touch-sensitive screen. Alternatively, the user can input the digits via the touchpad 116 or via the marked region 216 representing a touchpad on the touch-sensitive screen 110, in which case the input is converted into a digit by means of handwriting recognition. In FIG. 2, the user has already input the first three digits of the telephone number, 0, 1 and 2, and has just finished inputting the fourth digit 3. For this purpose, the user has drawn the digit 3 on the touchpad 116 (not illustrated) or on the corresponding region 216 of the touch-sensitive screen 110 using his finger or another suitable input device. In the figure, the digit input last is illustrated in a slightly enlarged manner in comparison with the previously input digits. This can be used to indicate the conclusion of the recognition. In order to adapt the recognition algorithm for handwriting recognition for the input of digits, regions of the screen 110 around the input field 202 were subjected to optical character recognition. For example, after optical character recognition, a first region 206 is examined for terms which indicate the type of characters or gestures to be input into the input field 202. The first region 206 is indicated by the dot-dashed line. In this case, the term call number is illustrated on the left beside the input field 202 in the first region 206. A comparison with corresponding terms in a database makes it possible to conclude that only a limited selection of alphanumeric characters can be input. For a telephone number, these are the digits 0-9, the star key and the hash key, for example. A second region 208 which includes the input field and is larger than the first region 206 is illustrated in the figure. The second region 208 is indicated by the dashed line. In addition to the term call number, optical character recognition carried out in this region also recognizes the term telephone which likewise indicates the type of alphanumeric characters to be input. The occurrence of two terms describing the characters to be input into the input field in the second region 208 can improve the determination of the type of input field and therefore the ultimate recognition of the characters to be input into the input field.

[0044] FIG. 3 shows a schematic flowchart of the method 300 according to the invention. In step 302, character recognition or voice recognition is carried out in at least one region of the screen which includes a selected or active input field. The result of the character recognition or voice recognition is used to determine a type of input field in step 304. In the optional step 308, depending on the type of input field, a so-called speller is searched for in the screen contents or a read connection to a database of the type in question is set up directly, locally or remotely, if possible. Alternatively, possible structure rules for the input are searched for locally or remotely. On the basis of the result of the determination of the type of input field, an adapted recognition algorithm for the handwriting recognition or an adapted parameter set for the recognition algorithm is selected. The figure does not illustrate the selection of the region and the transfer of the region to be evaluated by the optical character recognition to the character recognition function or apparatus.

[0045] FIG. 4 shows a schematic illustration of a block diagram of an apparatus according to the invention. First means 400 which are set up to carry out optical character recognition in a region of the screen which comprises at least the input field and the immediate environment of the input field are communicatively connected to second means 402 which are set up to search for terms describing field types in the result of the optical character recognition via one or more data buses 410. Instead of or in addition to the first means 400 for the optical character recognition, it is possible to provide further first means 404 which are set up to carry out voice recognition for a voice instruction acoustically output after the input field has been displayed. The further first means 404 are likewise communicatively connected to the second means via the one or more data buses 410. One or more screens 406 are likewise communicatively connected at least to the first means 400 via the one or more data buses 410 in order to make it possible for the first means 400 to access that region of the screen which comprises at least the input field and the immediate environment of the input field or to transmit the image information to the first means 400. Finally, third means 408 are communicatively connected at least to the second means 402 via the one or more data buses 410 in order to select a recognition algorithm which is adapted to the handwritten input of alphanumeric characters and/or gestures or an adapted parameter set for the recognition algorithm on the basis of the terms which describe field types and were found by the second means. The third means comprise, for example, a database and an apparatus or a computer program which compares the terms found using the character or voice recognition with terms stored in the database and, in the event of a hit, selects a corresponding recognition algorithm adapted to a field type or an adapted parameter set for the recognition algorithm.

[0046] Optionally, it is also possible to provide means 412 which dynamically control the input with respect to a database taking into account already available contents by restricting the inputs which are still expected. This can be carried out locally or remotely, for example, using speller functionalities which are already provided by the corresponding application or alternatively by directly accessing the corresponding database, or structure rules for the input are retrieved locally or remotely. These structure rules may be country-dependent, for example. This step is used to determine an extended parameter set for the recognition algorithm.