fix: 增强噪声压制 - 检测混合文字乱码(Latin+CJK交错)
- 在_is_noise_text中新增混合文字乱码检测 - 检测Latin大写字母被CJK分割成多个不连续块的模式 - 如'HHWH尸端自哪内柙登录槟J!'这类OCR垃圾现在会被过滤 - 修复unicode_radicals.md末尾格式错误(缺少补充字符列) - 从git追踪中移除easyocr_worker_state.json(运行时状态文件)
This commit is contained in:
+2
-2
@@ -218,8 +218,8 @@
|
||||
| 丿 | U+4E3F | 撇 | 儿→丿 | 儿 |
|
||||
| 丨 | U+4E28 | 竖 | 中→丨 | 中 |
|
||||
| 虱 | U+8671 | 虱 | 虱→搁 | 虱 |
|
||||
| 凵 | U+51F5 | 凵 | 凵→儿| 凵
|
||||
| 乍 | U+4E4D | 作 | 乍→作 | 乍
|
||||
| 凵 | U+51F5 | 凵 | 凵→儿 | 凵 |
|
||||
| 乍 | U+4E4D | 作 | 乍→作 | 乍 |
|
||||
|
||||
## 使用说明
|
||||
|
||||
|
||||
Reference in New Issue
Block a user