The current version used is com.google.mlkit:text-recognition-chinese:16.0.0-beta4, but it has a low accuracy. I've searched Google, but couldn't find a better solution. Any recommendations from fellow developers are welcome.