Help Center> Optical Character Recognition> API Reference> API> General Text

Updated on 2024-07-11 GMT+08:00

View PDF

General Text

Function

This API detects and extracts text from images and converts the text and coordinates into JSON format. It can be used in various scenarios, such as scanned files, electronic documents, books, receipts, and forms.

English and Chinese are supported but support for traditional Chinese characters is limited. For details about the constraints on using this API, see Constraints and Limitations. For details about how to use this API, see Introduction to OCR.

Constraints and Limitations

Only images in PNG, JPG, JPEG, BMP, GIF, TIFF, WebP, PCX, ICO, or PSD format can be recognized.
No side of the image can be smaller than 15 or larger than 8,192 pixels.
The area to be recognized must occupy more than 80% of the image. When scanning a table, ensure that all text and its surrounding area are included in the image.
An image can be rotated to any angle.
Light-colored text watermarks can be automatically filtered out.
Text in images with complex backgrounds (such as outdoor scenery) or distorted text cannot be recognized.
Supported languages: Chinese, English, some traditional Chinese, Malay, Ukrainian, Hindi, Russian, Vietnamese, Indonesian, Thai, Arabic, German, Latin, French, Italian, Spanish, Portuguese, Romanian, Polish Amharic, Japanese, Korean, Turkish, Norwegian, Danish, Swedish, and Khmer.

Calling Method

For details, see Calling APIs.

Prerequisites

Before using this API, subscribe to the service and complete authentication. For details, see Subscribing to an OCR Service and Authentication.

Before using the service for the first time, you need to enable the service by clicking Subscribe. You only need to subscribe to the service once. If you have not subscribed to the service yet, error "ModelArts.4204" will be displayed when you call this API. Before you call the API, log in to the OCR console and subscribe to the corresponding service. Ensure that you make the subscription to the service in the same region where you want to call this API.

URI

POST /v2/{project_id}/ocr/general-text

**Table 1** URI parameters
Parameter	Mandatory	Description
endpoint	Yes	Endpoint, which is the request address for calling an API. The endpoint varies depending on services in different regions. For more details, see Endpoints.
project_id	Yes	Project ID, which can be obtained from Obtaining a Project ID.

Request Parameters

**Table 2** Request header parameters
Parameter	Mandatory	Type	Description
X-Auth-Token	Yes	String	User token Used to obtain the permission to call APIs. The token is the value of X-Subject-Token in the response header in Authentication.
Content-Type	Yes	String	MIME type of the request body. The value is application/json.
Enterprise-Project-Id	No	String	Enterprise project ID. OCR uses Enterprise Project Management Service (EPS) to split fees for resources used by different user groups and users. To obtain the enterprise project ID, go to the Enterprise Project Management console, click the enterprise project name, and obtain the enterprise project ID on the enterprise project details page. For details about how to create an enterprise project, see Optical Character Recognition User Guide. NOTE: After an enterprise project is created, parameter transfer involves the following scenarios: If a correct enterprise project ID is carried and the OCR service can be used properly, the bills will be categorized under the corresponding enterprise project for that ID. If an enterprise project ID that is in the correct format but does not actually exist is carried, and the OCR service can be used properly, the bills will display the corresponding non-existent enterprise project ID. If no enterprise project ID or an enterprise project ID with incorrect format (such as special characters) is carried, and the OCR service can be used properly, the bills will be categorized under default.

**Table 3** Request body parameters
Parameter	Mandatory	Type	Description
image	No	String	Set either this parameter or url. Base64 encoded string of an image file. The image file has a size limit of 10 MB. No side of the image can be smaller than 15 or larger than 8,192 pixels. Only images in JPEG, JPG, PNG, BMP, GIF, TIFF, WebP, PCX, ICO, or PSD format can be recognized. An example is /9j/4AAQSkZJRgABAg.... If the image data contains an unnecessary prefix, the error "The image format is not supported" is reported.
url	No	String	Set either this parameter or image. Image URL. Currently, the following URLs are supported: Public HTTP/HTTPS URL URL provided by OBS. You need to be authorized to use OBS data, including service authorization, temporary authorization, and anonymous public authorization. For details, see Configuring Access Permissions of OBS. NOTE: The API response time depends on the image download time. If the image download takes a long time, the API call will fail. Ensure that the storage service where the image to be detected resides is stable and reliable. OBS is recommended for storing image data. The URL cannot contain Chinese characters. If Chinese characters exist, they must be encoded using UTF-8.
detect_direction	No	Boolean	Whether to align the tilted image. The options are: true: The tilted image will be aligned. false: The tilted image will not be aligned. An image tilted to any angle can be aligned. If this parameter is not specified, false is used by default. If the image to be recognized is tilted, you are advised to set this parameter to true.
quick_mode	No	Boolean	Whether to enable the quick mode. For a single-line text image (the image contains only one line of text and the text area occupies more than 50% of the image), the recognition results can be returned more quickly when this quick mode is enabled. The options are: true: The quick mode will be enabled. false: The quick mode will be disabled. If this parameter is not specified, false is used by default. In this case, the quick mode will be disabled.
character_mode	No	Boolean	Whether to enable the single-character mode. The options are: true: The single-character mode is enabled. false: The single-character mode is disabled. If this parameter is not transferred, the default value false is used, and information about a single character that occupies a text line is not returned.
language	No	String	Language. If this parameter is not specified, Chinese and English will be used by default. The options are: auto: automatic language classification ms: Malay uk: Ukrainian hi: Hindi ru: Russian vi: Vietnamese id: Indonesian th: Thai zh: Chinese and English ar: Arabic de: German la: Latin fr: French it: Italian es: Spanish pt: Portuguese ro: Romanian pl: Polish am: Amharic ja: Japanese ko: Korean tr: Turkish no: Norwegian da: Danish sv: Swedish km: Khmer he: Hebrew
single_orientation_mode	No	Boolean	Whether to enable the single direction mode. The options are: true: The single direction mode is enabled. false: The single direction mode is disabled. If this parameter is not specified, false is used by default. In this case, the fields in the image are recognized as in multiple directions by default.

Response Parameters

The status code may vary depending on the recognition results. For example, 200 indicates that the API is successfully called, and 400 indicates that the API fails to be called. The following describes the status codes and corresponding response parameters.

Status code: 200

**Table 4** Response body parameter
Parameter	Type	Description
result	GeneralTextResult object	Recognition result This parameter is not returned when the API fails to be called.

**Table 5** GeneralTextResult
Parameter	Type	Description
direction	Float	Image direction This parameter is valid only when detect_direction is set to true. The anti-clockwise rotation angle of an image is returned. The value ranges from 0 to 359. When detect_direction is set to false, the value of this parameter is -1.
words_block_count	Integer	Number of detected text blocks
words_block_list	Array of GeneralTextWordsBlockList objects	List of recognized text blocks. The output sequence is from left to right and from top to bottom.

**Table 6** GeneralTextWordsBlockList
Parameter	Type	Description
words	String	Recognition result of a text block
location	Array<Array<Integer>>	List of location information about a text block, including the 2D coordinates (x, y) of four vertexes in the text area, where the coordinate origin is the upper-left corner of the image, the X axis is horizontal, and the Y axis is vertical.
confidence	Float	Confidence of a recognized text block
char_list	Array of GeneralTextCharList objects	Single-character recognition list corresponding to a text block. The output sequence is from left to right and from top to bottom.

**Table 7** GeneralTextCharList
Parameter	Type	Description
char	String	Recognition result of a single character
char_location	Array<Array<Integer>>	List of location information about a single character, including the 2D coordinates (x, y) of four vertexes in the character area, where the coordinate origin is the upper-left corner of the image, the X axis is horizontal, and the Y axis is vertical.
char_confidence	Float	Confidence of a recognized character

Status code: 400

**Table 8** Response body parameters
Parameter	Type	Description
error_code	String	Error code of a failed API call. For details, see Error Codes. This parameter is not returned when the API is successfully called.
error_msg	String	Error message when the API call fails This parameter is not returned when the API is successfully called.

Example Request

endpoint is the request URL for calling an API. Endpoints vary depending on services and regions. For details, see Endpoints.
For example, General Text OCR is deployed in the CN-Hong Kong region. The endpoint is ocr.ap-southeast-1.myhuaweicloud.com or ocr.ap-southeast-1.myhuaweicloud.cn. The request URL is https://ocr.ap-southeast-1.myhuaweicloud.com/v2/{project_id}/ocr/general-text. project_id is the project ID. For how to obtain the project ID, see Obtaining a Project ID.
For details about how to obtain a token, see Authentication.

Transfer the Base64 encoded string of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

POST https://{endpoint}/v2/{project_id}/ocr/general-text
 Request Header:   
 Content-Type: application/json   
 X-Auth-Token: MIINRwYJKoZIhvcNAQcCoIINODCCDTQCAQExDTALBglghkgBZQMEAgEwgguVBgkqhkiG...   
 Request Body:
 {   
    "image":"/9j/4AAQSkZJRgABAgEASABIAAD/4RFZRXhpZgAATU0AKgAAAA...",
    "detect_direction":false,
    "quick_mode":false
  }

Transfer the URL of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

POST https://{endpoint}/v2/{project_id}/ocr/general-text
 Request Header:   
 Content-Type: application/json   
 X-Auth-Token: MIINRwYJKoZIhvcNAQcCoIINODCCDTQCAQExDTALBglghkgBZQMEAgEwgguVBgkqhkiG...   
 Request Body:
 {
     "url":"https://BucketName.obs.xxxx.com/ObjectName",
     "detect_direction":false,
     "quick_mode":false
  }

Example Response

Status code: 200

Example response for a successful request

{
  "result" : {
    "direction" : 67.6506,
    "words_block_count" : 1,
    "words_block_list" : [ {
      "words" : "Word",
      "confidence" : 0.9999,
      "location" : [ [ 517, 447 ], [ 540, 504 ], [ 505, 518 ], [ 482, 461 ] ],
      "char_list" : [ {
        "char" : "Character",
        "char_location" : [ [ 517, 447 ], [ 530, 479 ], [ 495, 493 ], [ 482, 461 ] ],
        "char_confidence" : 0.9999
      }, {
        "char" : "Character",
        "char_location" : [ [ 530, 479 ], [ 540, 504 ], [ 505, 518 ], [ 495, 493 ] ],
        "char_confidence" : 0.9999
      } ]
    } ]
  }
}

Status code: 400

Example response for a failed request

{
    "error_code": "AIS.0103",
    "error_msg": "The image size does not meet the requirements."
}

Example SDK Code

The example SDK code is as follows:

You are advised to update the SDKs to the latest versions before use to prevent the local outdated SDKs from being unable to use the latest OCR functions.

Transfer the Base64 encoded string of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

        
         
           
           
             package com.huaweicloud.sdk.test;

import com.huaweicloud.sdk.core.auth.ICredential;
import com.huaweicloud.sdk.core.auth.BasicCredentials;
import com.huaweicloud.sdk.core.exception.ConnectionException;
import com.huaweicloud.sdk.core.exception.RequestTimeoutException;
import com.huaweicloud.sdk.core.exception.ServiceResponseException;
import com.huaweicloud.sdk.ocr.v1.region.OcrRegion;
import com.huaweicloud.sdk.ocr.v1.*;
import com.huaweicloud.sdk.ocr.v1.model.*;


public class RecognizeGeneralTextSolution {

    public static void main(String[] args) {
        // The AK and SK used for authentication are hard-coded or stored in plaintext, which has great security risks. It is recommended that the AK and SK be stored in ciphertext in configuration files or environment variables and decrypted during use to ensure security.
        // In this example, AK and SK are stored in environment variables for authentication. Before running this example, set environment variables CLOUD_SDK_AK and CLOUD_SDK_SK in the local environment
        String ak = System.getenv("CLOUD_SDK_AK");
        String sk = System.getenv("CLOUD_SDK_SK");

        ICredential auth = new BasicCredentials()
                .withAk(ak)
                .withSk(sk);

        OcrClient client = OcrClient.newBuilder()
                .withCredential(auth)
                .withRegion(OcrRegion.valueOf("<YOUR REGION>"))
                .build();
        RecognizeGeneralTextRequest request = new RecognizeGeneralTextRequest();
        GeneralTextRequestBody body = new GeneralTextRequestBody();
        body.withQuickMode(false);
        body.withDetectDirection(false);
        body.withImage("/9j/4AAQSkZJRgABAgEASABIAAD/4RFZRXhpZgAATU0AKgAAAA...");
        request.withBody(body);
        try {
            RecognizeGeneralTextResponse response = client.recognizeGeneralText(request);
            System.out.println(response.toString());
        } catch (ConnectionException e) {
            e.printStackTrace();
        } catch (RequestTimeoutException e) {
            e.printStackTrace();
        } catch (ServiceResponseException e) {
            e.printStackTrace();
            System.out.println(e.getHttpStatusCode());
            System.out.println(e.getRequestId());
            System.out.println(e.getErrorCode());
            System.out.println(e.getErrorMsg());
        }
    }
}

            

          

        
       

Transfer the URL of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

        
         
           
           
             package com.huaweicloud.sdk.test;

import com.huaweicloud.sdk.core.auth.ICredential;
import com.huaweicloud.sdk.core.auth.BasicCredentials;
import com.huaweicloud.sdk.core.exception.ConnectionException;
import com.huaweicloud.sdk.core.exception.RequestTimeoutException;
import com.huaweicloud.sdk.core.exception.ServiceResponseException;
import com.huaweicloud.sdk.ocr.v1.region.OcrRegion;
import com.huaweicloud.sdk.ocr.v1.*;
import com.huaweicloud.sdk.ocr.v1.model.*;


public class RecognizeGeneralTextSolution {

    public static void main(String[] args) {
        // The AK and SK used for authentication are hard-coded or stored in plaintext, which has great security risks. It is recommended that the AK and SK be stored in ciphertext in configuration files or environment variables and decrypted during use to ensure security.
        // In this example, AK and SK are stored in environment variables for authentication. Before running this example, set environment variables CLOUD_SDK_AK and CLOUD_SDK_SK in the local environment
        String ak = System.getenv("CLOUD_SDK_AK");
        String sk = System.getenv("CLOUD_SDK_SK");

        ICredential auth = new BasicCredentials()
                .withAk(ak)
                .withSk(sk);

        OcrClient client = OcrClient.newBuilder()
                .withCredential(auth)
                .withRegion(OcrRegion.valueOf("<YOUR REGION>"))
                .build();
        RecognizeGeneralTextRequest request = new RecognizeGeneralTextRequest();
        GeneralTextRequestBody body = new GeneralTextRequestBody();
        body.withQuickMode(false);
        body.withDetectDirection(false);
        body.withUrl("https://BucketName.obs.myhuaweicloud.com/ObjectName");
        request.withBody(body);
        try {
            RecognizeGeneralTextResponse response = client.recognizeGeneralText(request);
            System.out.println(response.toString());
        } catch (ConnectionException e) {
            e.printStackTrace();
        } catch (RequestTimeoutException e) {
            e.printStackTrace();
        } catch (ServiceResponseException e) {
            e.printStackTrace();
            System.out.println(e.getHttpStatusCode());
            System.out.println(e.getRequestId());
            System.out.println(e.getErrorCode());
            System.out.println(e.getErrorMsg());
        }
    }
}

            

          

        
       

Transfer the Base64 encoded string of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

        
         
           
           
             # coding: utf-8

from huaweicloudsdkcore.auth.credentials import BasicCredentials
from huaweicloudsdkocr.v1.region.ocr_region import OcrRegion
from huaweicloudsdkcore.exceptions import exceptions
from huaweicloudsdkocr.v1 import *

if __name__ == "__main__":
    # The AK and SK used for authentication are hard-coded or stored in plaintext, which has great security risks. It is recommended that the AK and SK be stored in ciphertext in configuration files or environment variables and decrypted during use to ensure security.
    # In this example, AK and SK are stored in environment variables for authentication. Before running this example, set environment variables CLOUD_SDK_AK and CLOUD_SDK_SK in the local environment
    ak = os.getenv("CLOUD_SDK_AK")
    sk = os.getenv("CLOUD_SDK_SK")

    credentials = BasicCredentials(ak, sk) \

    client = OcrClient.new_builder() \
        .with_credentials(credentials) \
        .with_region(OcrRegion.value_of("<YOUR REGION>")) \
        .build()

    try:
        request = RecognizeGeneralTextRequest()
        request.body = GeneralTextRequestBody(
            quick_mode=False,
            detect_direction=False,
            image="/9j/4AAQSkZJRgABAgEASABIAAD/4RFZRXhpZgAATU0AKgAAAA..."
        )
        response = client.recognize_general_text(request)
        print(response)
    except exceptions.ClientRequestException as e:
        print(e.status_code)
        print(e.request_id)
        print(e.error_code)
        print(e.error_msg)

            

          

        
       

Transfer the URL of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

        
         
           
           
             # coding: utf-8

from huaweicloudsdkcore.auth.credentials import BasicCredentials
from huaweicloudsdkocr.v1.region.ocr_region import OcrRegion
from huaweicloudsdkcore.exceptions import exceptions
from huaweicloudsdkocr.v1 import *

if __name__ == "__main__":
    # The AK and SK used for authentication are hard-coded or stored in plaintext, which has great security risks. It is recommended that the AK and SK be stored in ciphertext in configuration files or environment variables and decrypted during use to ensure security.
    # In this example, AK and SK are stored in environment variables for authentication. Before running this example, set environment variables CLOUD_SDK_AK and CLOUD_SDK_SK in the local environment
    ak = os.getenv("CLOUD_SDK_AK")
    sk = os.getenv("CLOUD_SDK_SK")

    credentials = BasicCredentials(ak, sk) \

    client = OcrClient.new_builder() \
        .with_credentials(credentials) \
        .with_region(OcrRegion.value_of("<YOUR REGION>")) \
        .build()

    try:
        request = RecognizeGeneralTextRequest()
        request.body = GeneralTextRequestBody(
            quick_mode=False,
            detect_direction=False,
            url="https://BucketName.obs.myhuaweicloud.com/ObjectName"
        )
        response = client.recognize_general_text(request)
        print(response)
    except exceptions.ClientRequestException as e:
        print(e.status_code)
        print(e.request_id)
        print(e.error_code)
        print(e.error_msg)

            

          

        
       

Transfer the Base64 encoded string of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

        
         
           
           
             package main

import (
	"fmt"
	"github.com/huaweicloud/huaweicloud-sdk-go-v3/core/auth/basic"
    ocr "github.com/huaweicloud/huaweicloud-sdk-go-v3/services/ocr/v1"
	"github.com/huaweicloud/huaweicloud-sdk-go-v3/services/ocr/v1/model"
    region "github.com/huaweicloud/huaweicloud-sdk-go-v3/services/ocr/v1/region"
)

func main() {
    // The AK and SK used for authentication are hard-coded or stored in plaintext, which has great security risks. It is recommended that the AK and SK be stored in ciphertext in configuration files or environment variables and decrypted during use to ensure security.
    // In this example, AK and SK are stored in environment variables for authentication. Before running this example, set environment variables CLOUD_SDK_AK and CLOUD_SDK_SK in the local environment
    ak := os.Getenv("CLOUD_SDK_AK")
    sk := os.Getenv("CLOUD_SDK_SK")

    auth := basic.NewCredentialsBuilder().
        WithAk(ak).
        WithSk(sk).
        Build()

    client := ocr.NewOcrClient(
        ocr.OcrClientBuilder().
            WithRegion(region.ValueOf("<YOUR REGION>")).
            WithCredential(auth).
            Build())

    request := &model.RecognizeGeneralTextRequest{}
	quickModeGeneralTextRequestBody:= false
	detectDirectionGeneralTextRequestBody:= false
	imageGeneralTextRequestBody:= "/9j/4AAQSkZJRgABAgEASABIAAD/4RFZRXhpZgAATU0AKgAAAA..."
	request.Body = &model.GeneralTextRequestBody{
		QuickMode: &quickModeGeneralTextRequestBody,
		DetectDirection: &detectDirectionGeneralTextRequestBody,
		Image: &imageGeneralTextRequestBody,
	}
	response, err := client.RecognizeGeneralText(request)
	if err == nil {
        fmt.Printf("%+v\n", response)
    } else {
        fmt.Println(err)
    }
}

            

          

        
       

Transfer the URL of the image for recognition. During the recognition, the tilt angle of the image is not verified, and the quick mode is disabled.

        
         
           
           
             package main

import (
	"fmt"
	"github.com/huaweicloud/huaweicloud-sdk-go-v3/core/auth/basic"
    ocr "github.com/huaweicloud/huaweicloud-sdk-go-v3/services/ocr/v1"
	"github.com/huaweicloud/huaweicloud-sdk-go-v3/services/ocr/v1/model"
    region "github.com/huaweicloud/huaweicloud-sdk-go-v3/services/ocr/v1/region"
)

func main() {
    // The AK and SK used for authentication are hard-coded or stored in plaintext, which has great security risks. It is recommended that the AK and SK be stored in ciphertext in configuration files or environment variables and decrypted during use to ensure security.
    // In this example, AK and SK are stored in environment variables for authentication. Before running this example, set environment variables CLOUD_SDK_AK and CLOUD_SDK_SK in the local environment
    ak := os.Getenv("CLOUD_SDK_AK")
    sk := os.Getenv("CLOUD_SDK_SK")

    auth := basic.NewCredentialsBuilder().
        WithAk(ak).
        WithSk(sk).
        Build()

    client := ocr.NewOcrClient(
        ocr.OcrClientBuilder().
            WithRegion(region.ValueOf("<YOUR REGION>")).
            WithCredential(auth).
            Build())

    request := &model.RecognizeGeneralTextRequest{}
	quickModeGeneralTextRequestBody:= false
	detectDirectionGeneralTextRequestBody:= false
	urlGeneralTextRequestBody:= "https://BucketName.obs.myhuaweicloud.com/ObjectName"
	request.Body = &model.GeneralTextRequestBody{
		QuickMode: &quickModeGeneralTextRequestBody,
		DetectDirection: &detectDirectionGeneralTextRequestBody,
		Url: &urlGeneralTextRequestBody,
	}
	response, err := client.RecognizeGeneralText(request)
	if err == nil {
        fmt.Printf("%+v\n", response)
    } else {
        fmt.Println(err)
    }
}

            

          

        
       

For more SDK code examples in various programming languages, see the Sample Code tab on the right of the API Explorer page, which can automatically generate corresponding SDK code examples.