CN111381683A

CN111381683A - Photographing recognition input method and software

Info

Publication number: CN111381683A
Application number: CN201811620611.0A
Authority: CN
Inventors: 薛康泰华
Original assignee: Individual
Current assignee: Individual
Priority date: 2018-12-28
Filing date: 2018-12-28
Publication date: 2020-07-07

Abstract

The invention discloses a character input method and software based on a photographing function, which are mainly used for expanding the function of an input method, are particularly suitable for application scenes of a mobile terminal, and for example, a user uses dictionary APP software of the mobile terminal to look up new characters and new words, and specifically comprise the following steps: 1. the user selects a photographing identification input method, and software is started; 2. the software calls a photographing function to snap a photo containing words to be input and displayed in a photo area of the input interface; 3. a user touches or scratches a word to be input in the photo area by a finger, the software segments a word picture from the touch or scratch position and calls a character recognition module to recognize the word; 4. and displaying the recognition result by the software for the user to confirm. The invention expands the input method function of the mobile terminal or other terminals by photographing and identifying, and improves the convenience of character input.

Description

Photographing recognition input method and software

Technical Field

The invention relates to a photographing recognition input method and software, and belongs to the technical field of software.

Background

The input method software is one of basic software of other terminals such as a mobile terminal, a personal computer and the like, the existing input method software supports spelling input, handwriting input, voice input and the like, different input methods have different characteristics and have defects, for example, the spelling input method is not convenient enough, if the handwriting input method meets words which cannot be written, the handwriting input method cannot input, and if the voice input method meets words which cannot be read, the voice input method cannot input. The input method related by the invention adopts a photographing identification method, can complete input as long as photos containing words can be shot, is not only convenient, but also not limited by reading and writing problems, and is particularly suitable for application scenes of mobile terminals.

Disclosure of Invention

The purpose of the invention is as follows: the invention aims to provide a photographing recognition input method, which improves the convenience of input of a mobile terminal or other terminals.

The technical scheme is as follows: the key steps of the photographing recognition input method are shown in the attached figure 1 and are explained as follows:

(1) calling a photographing function after software is started, and snapping a photo containing words to be input by a user and displaying the photo in a photo area of an input interface;

(2) a user touches or scratches a word to be input in the photo area by a finger, the software segments a word picture from the touch or scratch position and calls a character recognition module to recognize the word;

(3) and displaying the recognition result by the software for the user to confirm.

Further perfecting the technical scheme, the software segments word pictures from the user touch positions, one of the schemes is a boundary search method, and the description is as follows:

(1) setting a split view frame with an initial size, such as 10 × 10 pixels, with the touch point as the center;

(2) firstly, starting from one side of a split view-finding frame, judging whether the side intersects with the word or not, if so, extending a pixel outwards, and circulating until the side does not intersect with the word, namely determining a boundary of the searched word;

(3) and (3) sequentially executing the step (2) on the other three edges of the split view-finding frame until the four boundaries of the words are determined.

Further, in order to enable the user to visually see the boundary searching process of the scheme so as to enable the user to intervene in time, the software displays the split view-finding frame on the photo in real time while dynamically adjusting the boundary of the split view-finding frame, when the user sees that the word to be input completely falls into the split view-finding frame, the finger can be released, and the software finishes searching. If the user does not release the fingers, the software fixes the upper and lower boundaries on the basis of the last search result, and executes the search algorithm again to search the left and right boundaries of the words until the user releases the fingers to confirm.

Further perfecting the technical scheme, the software divides the word picture from the touch position of the user, the scheme can also use a boundary search method, and the difference lies in the setting of the initial size of the divided viewing frame, which is explained as follows:

(1) setting a split view frame with an initial size by taking a touch and stroke starting point as a starting point, wherein the initial length of the split view frame is the length of a finger touch and stroke track, and the initial width is an empirical value, such as 10 pixels;

Similarly, the software also displays the dynamic split view box on the photo in real time to let the user confirm when the splitting process can be ended, which is not described herein again.

The above-mentioned boundary search method is one of the word segmentation schemes that can be used in the present invention, but this scheme is not a limitation for restricting the scope of the right of the present invention, and in terms of technical implementation, other segmentation schemes may be selected, or multiple segmentation schemes may be mutually verified to improve the reliability of segmentation.

Further perfecting the technical scheme, the software calls a character recognition module to recognize words in the segmented picture, the character recognition module is composed of a plurality of sub-modules, the sub-modules can be expanded, and each sub-module corresponds to one language. The recognition sub-module may be implemented using mature OCR program code, such as Tesserac OCR program code, or GOCR program code, and different languages may use different algorithms and programs, but which algorithms and programs are used is not intended as a limitation on the scope of the present invention.

The photographing recognition input method software designed according to the technical scheme at least comprises the following functional modules:

(1) and the photographing module is used for photographing a photo containing words to be input. After the input method software is started, the module is called to take a picture, and the picture is displayed in the picture area.

(2) And the character segmentation module is used for segmenting the picture containing the words to be input from the picture. In addition to the boundary search method-based segmentation scheme described in this specification, other segmentation schemes may also be employed. The text segmentation module is invoked when the user taps or strokes the photo with a finger.

(3) And the character recognition module is used for calling the character recognition module to recognize the words from the divided picture after the division is finished. The character recognition module is composed of a plurality of character recognition submodules, the submodules can be expanded, and each submodule corresponds to one language. The recognition sub-module may be implemented using sophisticated open source character recognition OCR program code, such as Tesserac OCR program code, or GOCR program code, with different languages possibly using different algorithms and programs.

The invention also provides application of the photographing identification input method in a mobile phone input method and infant literacy software.

Has the advantages that: the invention extends the input method by photographing and identifying, can finish input as long as photos containing words can be photographed, is not only convenient, but also not limited by the problem of reading and writing, and is particularly suitable for application scenes of mobile terminals. For example, the input method is used in the mobile terminal electronic dictionary APP, when a user encounters a new word in the process of reading an article, the user can directly take a picture of the reading object, touch or touch and stroke the new word in the picture with a finger, automatically recognize the word by the input method, and look up the dictionary; or the user can directly take a picture to identify and look up the dictionary when encountering unknown signboards or guideboards when traveling to a strange mirror, so that the use is very convenient.

Drawings

FIG. 1 is a step diagram of the photographing recognition input method according to the present invention.

In the attached figure 1, after the input method software is called, a photographing module is called to photograph and display a picture containing words to be input, a user touches or scratches the words to be input in the picture, the software calls a segmentation module to segment the words, after the segmentation is completed, the software calls an identification module to identify the words in the segmented picture, and the identification result is displayed for the user to confirm.

Detailed description of the preferred embodiments

The technical solution of the present invention is described in detail below, but the scope of the present invention is not limited to the embodiments.

Example 1: the invention is used for a mobile phone input method, and the photographing identification input method module is expanded in mobile phone input method software and consists of the following sub-modules:

(1) and (5) a photographing submodule. When the user opens the input method and selects the 'photographing recognition' input method, the photographing submodule is called to capture a photo containing words to be input and display the photo in the photo area of the input interface.

(2) And partitioning the sub-modules. When a user touches or strokes the words in the photo area with a finger, the segmentation sub-module is called, the new words are segmented by using a boundary search method, and segmented pictures are obtained.

(3) English word discernment submodule piece. This module is implemented with Tesserac OCR program code. After the software obtains the segmentation picture, the module is called to identify English words in the segmentation picture.

Example 2: the invention is used for mobile phone infant literacy APP software, and the photographing identification input module is arranged in the mobile phone infant literacy APP software to help infants to recognize characters. The input module is composed of the following sub-modules

(1) And (5) a photographing submodule. The infant opens literacy APP software, calls the photographing submodule, shoots the infant literacy card photo, and displays the photo in the photo area.

(2) And partitioning the sub-modules. When the infant touches the words in the photo area with a finger, the segmentation submodule is called, the words are segmented by using a boundary search method, and segmented pictures are obtained.

(3) English word discernment submodule piece. The module is implemented in GOCR program code. After the segmentation picture is obtained, the module is called to identify English words in the segmentation picture, and APP software explains word meanings, pronunciation, example sentences and the like to help the infant to recognize characters.

As noted above, while the present invention has been shown and described with reference to certain preferred embodiments, it is not to be construed as limited thereto. Various changes in form and detail may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. A shooting recognition input method and software are characterized in that: comprises the following steps of (a) carrying out,

(2) the user touches or strokes the words to be input in the photo area by a finger, and the software automatically identifies the words at the touch or stroke position;

2. The photographing recognition input method and software of claim 1, wherein: the software automatically identifies the words of the positions of the strokes or strokes, and comprises the following two steps,

(1) segmenting word pictures from the positions of the touch or touch and scratch;

(2) and calling a character recognition module to recognize the words.

3. The software of claim 2, segmenting word pictures from user swipe points, wherein: and the software displays the dynamic segmentation view-finding frame in real time by taking the touch point as a reference position until the view-finding frame frames the word to be input, and the user releases the hands to confirm that the software completes segmentation.

4. The software of claim 2, wherein the word picture is divided from the user swipe location: and the software displays the dynamic segmentation view-finding frame in real time by taking the finger touch and stroke action track as a reference until the view-finding frame frames the word to be input, and the user releases the hand to confirm that the software finishes segmentation.

5. The software real-time display dynamic split viewfinder of claims 3 and 4, wherein: the center position and the length and width adjustment of the dynamic viewfinder are controlled by a segmentation algorithm.

6. The software-invoked word recognition module of claim 2, wherein: the word recognition module supports multi-language switching.

7. A shooting recognition input method and software are characterized in that: the software comprises the following modules which are,

(1) the shooting module is used for capturing a photo containing words to be input;

(2) the word segmentation module dynamically adjusts and displays a view-finding frame in real time under the control of a segmentation algorithm by taking the touch or touch-and-scratch position of a finger of a user as a reference position to finish word segmentation;

(3) the character recognition module is composed of a plurality of character recognition submodules, each submodule corresponds to one language, and the submodules can be expanded.