Reading Text from Images Using C#
Introduction
By using Optical Character Recognition (OCR), you can detect and extract handwritten and printed text present in an image. The API works with different surfaces and backgrounds. To use OCR as a service, you have to get a subscription key from the Microsoft Cognitive Service portal. Steps of registration and obtaining the API keys are mentioned in my previous article, "Analyzing Image Content Programmatically: Using the Microsoft Cognitive Vision API."
In this post, I will demonstrate handwriting and text recognition by uploading a locally stored image consuming the Cognitive API.
Embedded Text Recognition Inside an Image
Step 1
Create a new ASP.NET Web application in Visual Studio. Add a new ASPX page named TextRecognition.aspx. Then, replace the HTML code of TextRecognition.aspx with the following code.
In the following ASPX code, I have created an ASP panel. Inside that panel, I have added an Image, TextBox, a file upload control, and Submit button. Locally browsed images with handwritten or printed text will be displayed in the image control. A textbox will display text present in the image after the Submit button is clicked.
- <%@ Page Language="C#" AutoEventWireup="true"
- CodeBehind="TextRecognition.aspx.cs"
- Inherits="TextRecognition.TextRecognition" %>
- <!DOCTYPE html>
- <html xmlns="http://www.w3.org/1999/xhtml">
- <head runat="server">
- <title>OCR Recognition</title>
- </head>
- <body>
- <form id="myform" runat="server">
- <div>
- </div>
- <asp:Panel ID="ImagePanel" runat="server"
- Height="375px">
- <asp:Image ID="MyImage" runat="server"
- Height="342px" Width="370px" />
- <asp:TextBox ID="MyTestBox" runat="server"
- Height="337px" Width="348px"></asp:TextBox>
- <br />
- <input id="MyFileUpload" type="file"
- runat="server" />
- <input id="btnSubmit" runat ="server"
- type="submit" value="Submit"
- onclick="btnSubmit_Click" />
- <br />
- </asp:Panel>
- </form>
- </body>
- </html>
Step 2
Add the following namespaces in your TextRecognition.aspx.cs code file.
- using System;
- using System.Collections.Generic;
- using System.IO;
- using System.Linq;
- using System.Web;
- using System.Web.UI;
- using System.Web.UI.WebControls;
- using System.Net.Http;
- using System.Net.Http.Headers;
- using System.Diagnostics;
Step 3
Add the following app setting entries in your Web.config file. I have used an existing subscription key generated long back; you have to add your own subscription key registered from Azure Portal. The APIuri parameter key value is the endpoint of the ODR cognitive service.
- <appSettings>
- <add key="RequestParameters"
- value="language=unk&detectOrientation =true"/>
- <add key="APIuri"
- value="https://westcentralus.api.cognitive.microsoft.com/
- vision/v1.0/ocr?"/>
- <add key="Subscription-Key"
- value="ce765f110a1e4a1c8eb5d2928a765c61"/>
- <add key ="Contenttypes" value="application/json"/>
- </appSettings>
Step 4
Now, add the following code in your TextRecognition.aspx.cs codebehind file. All static functions will return appSettings key values, as mentioned in Step 3. The BtnSubmit_Click event will occur once the Submit button is clicked. It will call the CallAPIforOCR async function. By using the GetByteArray function, the local image will be converted to bytes and that would be passed to Cognitive Services for recognition.
- public partial class TextRecognition : System.Web.UI.Page
- {
- static string responsetext;
- static string responsehandwritting;
- static string Subscriptionkey()
- {
- return System.Configuration.ConfigurationManager
- .AppSettings["Subscription-Key"];
- }
- static string RequestParameters()
- {
- return System.Configuration.ConfigurationManager
- .AppSettings["RequestParameters"];
- }
- static string ReadURI()
- {
- return System.Configuration.ConfigurationManager
- .AppSettings["APIuri"];
- }
- static string Contenttypes()
- {
- return System.Configuration.ConfigurationManager
- .AppSettings["Contenttypes"];
- }
- protected void btnSubmit_Click(object sender, EventArgs e)
- {
- string fileName = System.IO.Path.GetFileName
- (MyFileUpload.PostedFile.FileName);
- MyFileUpload.PostedFile.SaveAs(Server.MapPath
- ("~/images/" + fileName));
- MyImage.ImageUrl = "~/images/" + fileName;
- CallAPIforOCR("~/images/" + fileName);
- MyTestBox.Text = responsetext;
- }
- static byte[] GetByteArray(string LocalimageFilePath)
- {
- FileStream ImagefileStream = new
- FileStream(LocalimageFilePath, FileMode.Open,
- FileAccess.Read);
- BinaryReader ImagebinaryReader = new
- BinaryReader(ImagefileStream);
- return ImagebinaryReader.ReadBytes((int)
- ImagefileStream.Length);
- }
- // Optical Character Reader
- static async void CallAPIforOCR(string LocalimageFilePath)
- {
- var ComputerVisionAPIclient = new HttpClient();
- try {
- ComputerVisionAPIclient.DefaultRequestHeaders
- .Add("Ocp-Apim-Subscription- Key",
- Subscriptionkey());
- string requestParameters = RequestParameters();
- string APIuri = ReadURI() + requestParameters;
- HttpResponseMessage myresponse;
- byte[] byteData = GetByteArray(LocalimageFilePath);
- var content = new ByteArrayContent(byteData);
- content.Headers.ContentType = new
- MediaTypeHeaderValue(Contenttypes());
- myresponse = await
- ComputerVisionAPIclient.PostAsync
- (APIuri, content);
- myresponse.EnsureSuccessStatusCode();
- responsetext = await myresponse.Content
- .ReadAsStringAsync();
- }
- catch (Exception e)
- {
- EventLog.WriteEntry("Text Recognition Error",
- e.Message + "Trace" + e.StackTrace,
- EventLogEntryType.Error, 121, short.MaxValue);
- }
- }
Step 5
Now, set TextRecognition.aspx as the default page and execute the Web application. After the page is displayed, click the browser button and open an local image with printed text on it. Click the Submit button to see the output.
To show the demo, I have used the following image downloaded from the Internet. Successful execution of the Cognitive Services API call returns OCR, including text, a bounding box for regions, lines, and words. On the right side of the panel, you can see the embedded test displayed in the text box.

Figure 1: Output of OCR recognition
Following is the JSON response from Cognitive Services.
- {
- "TextAngle": 0.0,
- "Orientation": "NotDetected",
- "Language": "en",
- "Regions": [
- {
- "BoundingBox": "21,246,550,132",
- "Lines": [
- {
- "BoundingBox": "21,246,550,33",
- "Words": [
- {
- "BoundingBox": "21,247,85,32",
- "Text": "How"
- },
- {
- "BoundingBox": "118,246,140,33",
- "Text": "Sundar"
- },
- {
- "BoundingBox": "273,246,121,33",
- "Text": "Pichai"
- },
- {
- "BoundingBox": "410,247,161,32",
- "Text": "Became"
- }
- ]
- },
- {
- "BoundingBox": "39,292,509,33",
- "Words": [
- {
- "BoundingBox": "39,293,72,32",
- "Text": "The"
- },
- {
- "BoundingBox": "126,293,99,32",
- "Text": "Most"
- },
- {
- "BoundingBox": "240,292,172,33",
- "Text": "Powerful"
- },
- {
- "BoundingBox": "428,292,120,33",
- "Text": "Indian"
- }
- ]
- },
- {
- "BoundingBox": "155,338,294,40",
- "Words": [
- {
- "BoundingBox": "155,338,118,33",
- "Text": "Inside"
- },
- {
- "BoundingBox": "287,338,162,40",
- "Text": "Google?"
- }
- ]
- }
- ]
- }
- ]
- }
Recognize Handwriting
For handwriting recognition from text present in an image, I have used the same application, but you have to change the APIuri path to point to the correct endpoint and update the RequestParameters key value added in the previous step.
- <appSettings>
- <add key="RequestParameters" value="handwriting=true"/>
- <add key="APIuri"
- value="https://westcentralus.api.cognitive
- .microsoft.com/vision/v1.0/recognizeText?"/>
- <add key="Subscription-Key"
- value="ce765f110a1e4a1c8eb5d2928a765c61"/>
- <add key ="Contenttypes" value="application/json"/>
- </appSettings>
Also, add the following ReadHandwritttingFromImage async method. This function will replace the CallAPIforOCR function call present in the btnSubmit_Click event.
- static async void ReadHandwritttingFromImage(string LocalimageFilePath)
- {
- HttpResponseMessage myresponse = null;
- IEnumerable<string> myresponseValues = null;
- string operationLocation = null;
- var ComputerVisionAPIclient = new HttpClient();
- ComputerVisionAPIclient.DefaultRequestHeaders.Add
- ("Ocp-Apim-Subscription-Key", Subscriptionkey());
- string requestParameters = RequestParameters();
- string APIuri = ReadURI() + requestParameters;
- byte[] byteData = GetByteArray(LocalimageFilePath);
- var content = new ByteArrayContent(byteData);
- content.Headers.ContentType = new
- MediaTypeHeaderValue(Contenttypes());
- try
- {
- myresponse = await ComputerVisionAPIclient
- .PostAsync(APIuri, content);
- myresponseValues = myresponse.Headers
- .GetValues("Operation-Location");
- }
- catch (Exception e)
- {
- EventLog.WriteEntry("Handwritting Recognition Error",
- e.Message + "Trace" + e.StackTrace,
- EventLogEntryType.Error, 121, short.MaxValue);
- }
- foreach (var value in myresponseValues)
- {
- operationLocation = value;
- break;
- }
- try
- {
- myresponse = await ComputerVisionAPIclient
- .GetAsync(operationLocation);
- responsehandwritting = await myresponse.Content
- .ReadAsStringAsync();
- }
- catch (Exception e)
- {
- EventLog.WriteEntry("Handwritting Recognition Error",
- e.Message + "Trace" + e.StackTrace,
- EventLogEntryType.Error, 121, short.MaxValue);
- }
- }
Now, execute the Web application again. After the page is displayed, click the browser button and open an local image with handwritten text on it. Click the Submit button to see the output.

Figure 2: Output of handwriting recognition
Reading Text from Images Using C#的更多相关文章
- Suspended Animation——《The Economist》阅读积累(考研英语二·2010 Reading Text 1)
[知识小百科] Damien Hirst(达米恩●赫斯特):生于1965年,是新一代英国艺术家的主要代表人物之一.他主导了90年代英国艺术发展并享有很高的国际声誉.赫斯特在1986年9月就读于伦敦大学 ...
- read content in a text file in python
** read text file in pythoncapability: reading =text= from a text file 1. open the IDLE text editor ...
- 【深度学习Deep Learning】资料大全
最近在学深度学习相关的东西,在网上搜集到了一些不错的资料,现在汇总一下: Free Online Books by Yoshua Bengio, Ian Goodfellow and Aaron C ...
- (转) Awesome - Most Cited Deep Learning Papers
转自:https://github.com/terryum/awesome-deep-learning-papers Awesome - Most Cited Deep Learning Papers ...
- python3.4 build in functions from 官方文档 翻译中
2. Built-in Functions https://docs.python.org/3.4/library/functions.html?highlight=file The Python i ...
- ios异常(crash)输出
最近突然想起友盟的sdk附带的一个功能:将闪退异常情况上报服务器,(stackflow,github)找了一些资料,自己写了一个demo,想起来好久没有写过blog了,顺便分享. 其实不止是ios,a ...
- python3的文件操作
open的原型定义在bultin.py中,是一种内建函数,用于处理文件 open(file, mode='r', buffering=None, encoding=None, errors=None, ...
- python27读书笔记0.3
#-*- coding:utf-8 -*- ##D.has_key(k): A predicate that returns True if D has a key k.##D.items(): Re ...
- LOAD DATA INFILE Syntax--官方
LOAD DATA [LOW_PRIORITY | CONCURRENT] [LOCAL] INFILE 'file_name' [REPLACE | IGNORE] INTO TABLE tbl_n ...
随机推荐
- Java NIO4:缓冲区Buffer(续)
一.什么是缓冲区 一个缓冲区对象是固定数量的数据的容器,其作用是一个存储器,或者分段运输区,在这里数据可被存储并在之后用于检索.缓冲区像前篇文章讨论的那样被写满和释放,对于每个非布尔原始数据 ...
- Elicpse使用技巧-打开选中文件文件夹或者包的当前目录
很多时候,我们需要在eclipse那里打开选中文件(文件夹,包)的当前目录,在资源管理器那里显示这个目录,这个时候,我们又不想采用“选中文件/文件夹/包名--右击--Properties--Locat ...
- rabbitMQ教程(五)rabbitmq 指令 以及解决web管理界面无法使用guest用户登录
安装最新版本的rabbitmq(3.3.1),并启用management plugin后,使用默认的账号guest登陆管理控制台,却提示登陆失败. 翻看官方的release文档后,得知由于账号gues ...
- 基于 Django的Ajax实现 文件上传
---------------------------------------------------------------遇到困难的时候,勇敢一点,找同学朋友帮忙,找导师求助. Ajax Ajax ...
- windows开机启动bat文件
1.运行 shell:startup 命令,如下: 2.创建bat的快捷方式,把改快捷方式添加到,C:\ProgramData\Microsoft\Windows\Start Menu\Program ...
- VS2012添加数据库连接时报错,未能加载文件或程序集microsoft.sqlserver.management.sdk.sfc
今天在VS2012中添加数据库连接时报错 未能加载文件或程序集microsoft.sqlserver.management.sdk.sfc,Version=11.0 查了很多资料,最后下载安装了Sha ...
- B. School Marks(典型贪心)
链接 [https://codeforces.com/contest/540/problem/B] 题意 某个人有n门成绩,k门已知,剩下的他可以个瞎改,但有个要求,最后分数和不超过x,且每门成绩不超 ...
- 通过C#调用,实现js加密代码的反混淆,并运行js函数
前一篇我测试了vba调用htmlfile做反混淆,并执行js加密函数的代码.本文换成C#实现. 联系QQ:564955427 C#操作JS函数,可以通过ScriptControl组件,但这个组件只能在 ...
- Hive基础
一.常用语句 二.嵌套语句 以上两句的查询结果相同. 三.关键字查询
- Ubuntu18.04更新源
一.备份/etc/apt/sources.list文件 cd /etc/apt sudo cp sources.list sources.list.old 二.选择国内常用的源 #阿里源 deb ht ...
