python手写多个字母识别_一个带界面的CNN手写数字识别，使用Python（tensorflow, kivy）实现...

最新推荐文章于 2022-10-16 11:51:40 发布

weixin_39837727

最新推荐文章于 2022-10-16 11:51:40 发布

阅读量464

点赞数

文章标签： python手写多个字母识别

本文链接：https://blog.csdn.net/weixin_39837727/article/details/111438379

版权

一个使用Python、Tensorflow和Kivy实现的手写数字识别应用，支持鼠标或触屏输入，能识别0-9的数字。通过Tensorflow训练CNN模型，利用MNIST数据集，并用Kivy创建用户界面。

摘要由CSDN通过智能技术生成

CNN_Handwritten_Digit_Recognizer (CNN手写数字识别)

A CNN handwritten digit recognizer with graphical UI, implemented with Python (tensorflow, kivy).

一个带界面的CNN手写数字识别，使用Python(tensorflow, kivy)实现。

Introduction (简介)

This is a simple handwritten digit recognizer application, implemented with only Python language. Basically, I use Tensorflow to build the CNN classifier as the recognizer, and use Kivy to realize the UI. This program supports mouse or touch screen, and can recognize 0-9 altogether 10 numbers. The keypoint is as follows:

Use Tensorflow (as well as its Estimator API) to build the CNN model. Use MNIST dataset to train the model.

Use Kivy to build the UI.

After drawing digit on the Kivy App window, the app export the digit as a image, which, after being processed with PIL, is feeded to the CNN model. The recognition result is displayed o