手写数字识别 [英] handwritten digit recognition

问题描述

我想实现一个用于识别手写数字(波斯数字)的应用程序，并且我想知道哪种神经网络更适合实现它.我在获取输入时没有任何问题，并且我实现了该部分，并制作了8 * 8网格，该网格能够识别写入的正方形和完全空白的正方形！
我已经阅读了几篇有关该主题的文章，其中大多数都将反向传播和MLP用于他们的网络.由于我没有足够的时间来从事这个项目，我几乎为选择最佳而感到困惑！ />
如果有人帮助我并给我一些建议，我将不胜感激！

最好的问候

I want to implement an application for recognizing handwritten digits(persian digits),and I want to know that which type of neural network is more suitable for implementation of it.I don''t have any problem in getting inputs,and I implemented that part and made 8*8 grid that is able to recognize which square is written and which square is completely empty!
I have already read several articles about this subject and most of them have used backpropagation and MLP for their network.I''ve got almost confused for choosing the best because I don''t have enough time for this project!!!

I would appreciate if someone help me on it and give me some advice!

best regards

推荐答案

我在CP中发现了很多这样的东西.

使用内核判别分析的笔迹识别 [再谈手写识别:内核支持向量机 [用于识别C#中手写数字的神经网络 [用于识别手写数字的神经网络 [ GPU上的神经网络 [使用链码的光学字符识别(OCR)中的AC#项目 [ ^ ]

卷积神经网络MNIST工作台 [ Unicode光学字符识别 [

I have found this nice bunch in CP.

Handwriting Recognition using Kernel Discriminant Analysis[^]

Handwriting Recognition Revisited: Kernel Support Vector Machines[^]

Neural Network for Recognition of Handwritten Digits in C#[^]

Neural Network for Recognition of Handwritten Digits[^]

A Neural Network on GPU[^]

A C# Project in Optical Character Recognition (OCR) Using Chain Code[^]

Convolutional Neural Network MNIST Workbench[^]

Unicode Optical Character Recognition[^]

BTW Thanks for your question....

此 [ ^ ]可能会有所帮助.

This[^] might help.

这篇关于手写数字识别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！