最好的.NET API OCR？ [英] The best .NET OCR API?

问题描述

有人能列举一些C＃OCR的API？或仍是唯一的选择C ++与包装？

Could someone list some C# OCR APIs? Or is the only option still C++ with wrapper?

推荐答案

任何像样的OCR引擎需要数年时间来开发和大多数已在C或C ++开发。大多数工具包，如飞马，Asprise设有接口，C，C ++和C＃。许多供应商创建精简的包装是现有的OCR引擎，如ABBYY FineReader中，谷歌的tesseract和ScanSoft公司Caere（OMNIPAGE）拿去卖作为贴牌产品。

Any decent OCR engine takes years to develop and most have been developed in C or C++. Most toolkits such as Pegasus, Asprise are provided with interfaces to C, C++ and C# . Many vendors create thin wrappers are existing OCR engines such as Abbyy FineReader, Google Tesseract and Scansoft Caere (Omnipage) and sell them as a rebadged offering.

我不相信有在C＃或其他.NET语言编写纯粹出于商业级的OCR引擎。有可能是利用神经网络可能只能读取印在干净的文档字体的数量有限，在C＃编写一些不太复杂的OCR引擎。

I don't believe there are any commercial grade OCR engines written purely in C# or other .NET language. There may be some less complex OCR engines written in C# using neural networks that may only read a limited number of fonts printed on clean documents.

一个很好的商业级的OCR引擎结合许多不同的策略，以获得最佳的结果，包括图像预处理，字典，卦，多引擎和大量的训练数据，具有多年的测试。速度也可能成为较C或C ++与本机.NET版本的问题。

A good commercial grade OCR engine combines many different strategies to obtain the best results including image pre-processing, dictionaries, trigrams, multiple engines and large amounts of training data with many years of testing. Speed may also become an issue with a native .NET version compared to C or C++.

这可能是一个，而开发一个C＃引擎之前。移植现有的C ++代码可能被证明是非常困难的，可能是更容易从头开始。这是基于多年与许多不同的商业引擎的工作我个人的看法。

It may be a while before a C# engine is developed. Porting existing C++ code may prove to be very difficult and is probably easier to start from scratch. This is my personal opinion based on many years of working with many different commercial engines.

这篇关于最好的.NET API OCR？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！