使用拒绝类

<< 点击显示目录 >>

主页  mappVision帮助助手 > 机器视觉帮助 > mapp Vision  > 用例 > OCR - 识别和读取字符
> 预训练 OCR 字体
> 带正则化权重和拒绝级别的预定义字体
 >

使用拒绝类

每种字体都有带剔除类别(_Rej)和不带剔除类别(_NoRej)的两种版本。

如果字体使用了拒绝类别,这意味着找到的区域(样本)并不总是必须分配给一个字符,也可以分配给拒绝类别 "无字符"。如果训练字体中没有剔除类,则样本始终会被分配给一个字符(分级值最高的字符)。

在 OCR 中使用带有剔除类别的字体会影响生成最终结果的内部决策。因此,根据使用的字体是否带有剔除类别,结果也会相应不同。也不能排除字符串不同的情况。相同文本的 分级值 结果也可能不同。


Each font is available in a variant with a rejection class (_Rej) and in a variant without a rejection class (_NoRej).

If a rejection class is used for a font, this means that regions (samples) found do not always have to be assigned to a character, but can also be assigned to rejection class "No character". If there is no rejection class in the trained font, a sample is always assigned a character (the one with the highest GradingValue).

The use of a font with rejection class in OCR affects the internal decisions for generating the final result. The result may therefore differ accordingly depending on whether a font with or without a rejection class was used. It also cannot be ruled out that the character string differs. GradingValue with the same text can also turn out differently.