Tensorflow 之 chinese_hand_writing # 1 分析数据
in Tensorflow with 0 comment

Tensorflow 之 chinese_hand_writing # 1 分析数据

in Tensorflow with 0 comment

识别手写汉字要把识别手写洋文难上很多。

首先,英文字符的分类少,总共10+26*2;而中文总共50,000多汉字,常用的就有3000多。其次,汉字有书法,每个人书写风格多样。

数据来源

手写汉字数据集: CASIA-HWDB
下载HWDB1.1数据集:

wget http://www.nlpr.ia.ac.cn/databases/download/feature_data/HWDB1.1trn_gnt.zip
# zip解压没得说, 之后还要解压alz压缩文件
wget http://www.nlpr.ia.ac.cn/databases/download/feature_data/HWDB1.1tst_gnt.zip

注意HWDB1.1trn_gnt.zip解压出来的alz还要继续解压,这个alz真是难以捉摸。。。

分析数据

import os
import numpy as np
import struct
 
train_data_dir = "HWDB1.1trn_gnt"
test_data_dir = "HWDB1.1tst_gnt"
 
# 读取图像和对应的汉字
def read_from_gnt_dir(gnt_dir=train_data_dir):
    def one_file(f):
        header_size = 10
        while True:
            header = np.fromfile(f, dtype='uint8', count=header_size)
            if not header.size: break
            sample_size = header[0] + (header[1]<<8) + (header[2]<<16) + (header[3]<<24)
            tagcode = header[5] + (header[4]<<8)
            width = header[6] + (header[7]<<8)
            height = header[8] + (header[9]<<8)
            if header_size + width*height != sample_size:
                break
            image = np.fromfile(f, dtype='uint8', count=width*height).reshape((height, width))
            yield image, tagcode
 
    for file_name in os.listdir(gnt_dir):
        if file_name.endswith('.gnt'):
            file_path = os.path.join(gnt_dir, file_name)
            with open(file_path, 'rb') as f:
                for image, tagcode in one_file(f):
                    yield image, tagcode
 
# 统计样本数
train_counter = 0
test_counter = 0
for image, tagcode in read_from_gnt_dir(gnt_dir=train_data_dir):
    tagcode_unicode = struct.pack('>H', tagcode).decode('gb2312')
    train_counter += 1
for image, tagcode in read_from_gnt_dir(gnt_dir=test_data_dir):
    tagcode_unicode = struct.pack('>H', tagcode).decode('gb2312')
    test_counter += 1
 
# 样本数
print(train_counter, test_counter)

509733 223991

训练集:509733,测试集:223991

把刚才的image和tagcode_unicode拎回来

%matplotlib inline
import matplotlib.pyplot as plt
plt.imshow(255-image,interpolation='nearest',cmap='binary')
xlabel = "True: {0}".format(tagcode_unicode)
plt.xlabel(xlabel, fontproperties='SimHei')
plt.show()

mark

本来是黑底白字,实在看不下去。。。用255-image换了一下颜色

参考

Responses

From now on, bravely dream and run toward that dream.
陕ICP备17001447号