利用神经网络实现手写字识别

神经网络介绍

神经网络即多层感知机

如果不知道感知机的可以看博主之前的文章感知机及Python实现

神经网络实现及手写字识别

关于数据集：

从http://yann.lecun.com/exdb/mnist/下载，下载后将文件解压
将main函数中的path改为下载文件的存储路径即可

如果对数据集有问题，可以私信博主

关于实现：

基于pytorch实现，包括神经网络的构建，激活函数的选择
归一化使用了像素值/255的方式实现，可以尝试用别的方式进行归一化处理

import numpy as np
import torch
import torch.nn.functional as F
import torch.optim as optim
from torch import nn
class Data:'''this class is about data module'''def __init__(self):self.start_loc_image=16 # the start location of image dataself.start_loc_label=8 # the start location of label dataself.num_pixel=28*28 # the number of pixelsself.choice={'train-image':'train-images.idx3-ubyte','train-label':'train-labels.idx1-ubyte','test-image':'t10k-images.idx3-ubyte','test-label':'t10k-labels.idx1-ubyte'} # the specific file namedef get(self,path,train_test='',image_label=''):'''get the data from the file whose path is "path":param path: the saving path of given files, default is "./file/":param train_test: the data category("train" or "test"):param image_label: the data information("image" or "label"):return: the data you want'''if train_test not in ['train','test'] or image_label not in ['image','label']:raise NameError('please check you spelling,"train_test" can be "train/test", "image_label" can be "image/label"')ch=train_test+'-'+image_labeldata=[]if image_label=='image':print('loading images ...')with open(path+self.choice[ch],'rb',) as f:file=f.read()for i in range(self.start_loc_image,file.__len__(),self.num_pixel):item=[]pixel=file[i:i+self.num_pixel].hex()for p in range(0,pixel.__len__(),2):item.append(int(pixel[p:p+2],16)) # decode -> get the pixel information from original filedata.append(self.transform2image(item))f.close()elif image_label=='label':print('load labels ...')with open(path+self.choice[ch],'rb',) as f:file=f.read()for i in range(self.start_loc_label,file.__len__()):data.append(file[i]) # decode -> get the label from original filef.close()return datadef transform2image(self,data:list):'''transform pixel point to image:param data: the original 1D pixel points:return: transformed image(28*28)'''assert data.__len__()==784import numpy as npreturn np.reshape(data,(28,-1))def transfer_tensor(self,data):'''transfer data to tensor format:param data: the original input data:return: transferred data'''return torch.tensor(data)def normalize(self,data,maximum=255):'''normalize the data with maximum:param data: the input data:param maximum: the maximum of pixel(is 255):return: normalized data'''return torch.div(data,maximum)
class Network(nn.Module):'''this class is about neural network'''def __init__(self,in_dim,n_hidden,out_dim):'''define the network:param in_dim: the input dimension:param n_hidden: the hidden layer dimension:param out_dim: the output dimension'''### about networksuper(Network, self).__init__()self.layer1 = nn.Sequential(nn.Linear(in_dim,n_hidden))self.layer2 = nn.Sequential(nn.Linear(n_hidden,out_dim))### other parameterself.size_pixel=28*28 # the size of pixed of each picture### learning rateself.learning_rate=0.001### optimizerself.optimizer=optim.SGD(self.parameters(),lr=self.learning_rate)def forward(self,data):'''forward the input data:param data: the data you want to train:return: the output or predicted value'''hidden=F.relu(self.layer1(data))out=F.sigmoid(self.layer2(hidden))return outdef accuracy(self,act,pre):'''calculate the accuracy:param act: actual value:param pre: predicted value:return: accuracy'''assert act.__len__()==pre.__len__()return round((act==pre).sum().item()/act.__len__(),3)def pre_process(self,feature,label):'''pre processing:param feature: feature:param label: label:return: preprocessed feature and label'''### transform to the format of tensorfeature=dat.transfer_tensor(feature)feature=dat.normalize(feature)feature=feature.view(-1,self.size_pixel)label=dat.transfer_tensor(label)return feature,torch.tensor(label,dtype=torch.int64)def per_train(self,epoch,feature,label,validation=0.2,batch=50,verbose=True,num_view=50):'''train neural network:param epoch: training times:param feature: feature:param label: label:param validation: for using evaluation:param batch: batch size:param verbose: whether view the training process or not:param num_view: view via training "num_view" times:return: none'''assert feature.__len__()==label.__len__()print('training neural network ...')fea,lab=self.pre_process(feature,label)len_train=int(feature.__len__()*(1-validation))data_train,label_train=fea[:len_train+1],lab[:len_train+1]data_train=[data_train[i:i+batch]for i in range(0,len_train,batch)]label_train=[label_train[i:i+batch]for i in range(0,len_train,batch)]data_val,label_val=fea[len_train:],lab[len_train:]for e in range(epoch+1):self.train()loss_tmp=[]for img,lab in zip(data_train,label_train):pre=self(img)loss_train=nn.CrossEntropyLoss()(pre,lab)loss_tmp.append(loss_train)loss_train.backward()self.optimizer.step()self.optimizer.zero_grad()if verbose and e>0 and e%num_view==0:self.eval()pre=self(data_val)loss_val=nn.CrossEntropyLoss()(pre,label_val)_,pre_view=pre.max(1)acc=self.accuracy(label_val,pre_view)print('epoch: '+str(e)+'/'+str(epoch)+' --> training loss:',loss_train.item(),'validation loss:',loss_val.item(),'validation accuracy:',acc)
def set_seed(seed):'''set random seed in order that result can be replayed:param seed: random seed:return: none'''import randomtorch.manual_seed(seed)torch.cuda.manual_seed(seed)torch.backends.cudnn.deterministic = Truenp.random.seed(seed)random.seed(seed)
Seed=0
Size_pixel= 28*28
Hidden=200
Output=10
set_seed(Seed)
if __name__ == '__main__':### set some necessary parameterspath='./file/' # the path of saved file### initialize necessary classdat=Data() # for data relatednet=Network(Size_pixel,Hidden,Output) # build network### load image and label of training data and testing data# train_image=dat.get(path,'train','image')# train_label=dat.get(path,'train','label')test_image=dat.get(path,'test','image')test_label=dat.get(path,'test','label')### train networkepoch=1000 # training timesnet.per_train(epoch,test_image,test_label)

结果展示

可以看到随着训练次数的增加，预测手写字的准确率也越来越高
运行结果

loading images ...
load labels ...
training neural network ...
epoch: 50/1000 --> training loss: 2.1074414253234863 validation loss: 2.1338484287261963 validation accuracy: 0.716
epoch: 100/1000 --> training loss: 1.8810529708862305 validation loss: 1.9262853860855103 validation accuracy: 0.814
epoch: 150/1000 --> training loss: 1.7580647468566895 validation loss: 1.811558723449707 validation accuracy: 0.844
epoch: 200/1000 --> training loss: 1.6905272006988525 validation loss: 1.7511157989501953 validation accuracy: 0.858
epoch: 250/1000 --> training loss: 1.6461492776870728 validation loss: 1.7131295204162598 validation accuracy: 0.87
epoch: 300/1000 --> training loss: 1.6150131225585938 validation loss: 1.6869707107543945 validation accuracy: 0.877
epoch: 350/1000 --> training loss: 1.5923327207565308 validation loss: 1.6678192615509033 validation accuracy: 0.886
epoch: 400/1000 --> training loss: 1.5752110481262207 validation loss: 1.653057336807251 validation accuracy: 0.892
epoch: 450/1000 --> training loss: 1.5620046854019165 validation loss: 1.6411659717559814 validation accuracy: 0.896
epoch: 500/1000 --> training loss: 1.5515371561050415 validation loss: 1.6312130689620972 validation accuracy: 0.897
epoch: 550/1000 --> training loss: 1.5430898666381836 validation loss: 1.6226950883865356 validation accuracy: 0.903
epoch: 600/1000 --> training loss: 1.5360981225967407 validation loss: 1.615297794342041 validation accuracy: 0.905
epoch: 650/1000 --> training loss: 1.5302610397338867 validation loss: 1.608830451965332 validation accuracy: 0.907
epoch: 700/1000 --> training loss: 1.5252583026885986 validation loss: 1.603135108947754 validation accuracy: 0.908
epoch: 750/1000 --> training loss: 1.5209182500839233 validation loss: 1.5980955362319946 validation accuracy: 0.91
epoch: 800/1000 --> training loss: 1.5170769691467285 validation loss: 1.5936044454574585 validation accuracy: 0.911
epoch: 850/1000 --> training loss: 1.513597846031189 validation loss: 1.5895581245422363 validation accuracy: 0.915
epoch: 900/1000 --> training loss: 1.5103877782821655 validation loss: 1.5858769416809082 validation accuracy: 0.917
epoch: 950/1000 --> training loss: 1.5073943138122559 validation loss: 1.5824891328811646 validation accuracy: 0.918
epoch: 1000/1000 --> training loss: 1.5045477151870728 validation loss: 1.5793448686599731 validation accuracy: 0.92

效果展示

创作不易，觉得不错就微信扫码奖励一下吧！

利用神经网络实现手写字识别相关推荐

利用卷积神经网络实现手写字识别
本文我们介绍一下卷积神经网络,然后基于pytorch实现一个卷积神经网络,并实现手写字识别卷积神经网络介绍传统神经网络处理图片问题的不足让我们先复习一下神经网络的工作流程: 搭建一个神经网络将 ...
Pytorch入门练习2-kaggle手写字识别神经网络(CNN)实现
目录数据预处理自定义数据集构建网络结构对卷积神经网络进行训练和评估对数据进行预测保存预测数据,提交代码 SNN由于无法考虑到图片数据的维度关系,在预测精度上会被限制,本章我们采用CNN卷积 ...
Pytorch入门练习-kaggle手写字识别神经网络(SNN)实现
采用pytorch搭建神经网络,解决kaggle平台手写字识别问题. 数据来源:https://www.kaggle.com/competitions/digit-recognizer/data 参考 ...
matlab深度学习——【卷积神经网络】手写字的识别
> 本文所使用的数据集在文章最后,不需要积分就可以下载! 数据集下载这里主要是基于卷积神经网络的手写字的识别,我是用matlab做的,如果有对卷积神经网络不太熟悉的伙伴可以搜下,网上资源比较多 ...
基于tensorflow的MNIST手写字识别
一.卷积神经网络模型知识要点卷积卷积 1.卷积 2.池化 3.全连接 4.梯度下降法 5.softmax 本次就是用最简单的方法给大家讲解这些概念,因为具体的各种论文网上都有,连推导都有,所以本文主要 ...
TensorFlow基于minist数据集实现手写字识别实战的三个模型
手写字识别 model1:输入层→全连接→输出层softmax model2:输入层→全连接→隐含层→全连接→输出层softmax model3:输入层→卷积层1→卷积层2→全连接→dropout层→ ...
python手写汉字识别_TensorFlow 2.0实践之中文手写字识别
问题导读: 1.相比于简单minist识别,汉字识别具有哪些难点? 2.如何快速的构建一个OCR网络模型? 3.读取的时候有哪些点需要注意? 4.如何让模型更简单的收敛? 还在玩minist?fash ...
.net 数字转汉字_TensorFlow 2.0 中文手写字识别（汉字OCR）
TensorFlow 2.0 中文手写字识别(汉字OCR) 在开始之前,必须要说明的是,本教程完全基于TensorFlow2.0 接口编写,请误与其他古老的教程混为一谈,本教程除了手把手教大家完成这个 ...
在Windows上调试TensorFlow 2.0 中文手写字识别（汉字OCR）
在Windows上调试TensorFlow 2.0 中文手写字识别(汉字OCR) 一.环境的搭建 Windows+1080Ti+Cuda10.1 Tsorflow2.0.0 Numpy1.16.4 注 ...

利用神经网络实现手写字识别

神经网络介绍

神经网络实现及手写字识别

结果展示

利用神经网络实现手写字识别相关推荐

最新文章

热门文章