想问一下只OCR的话，文件名应该是什么样子的呢？

[TobisLee]

我有一个sup字幕文件，提取出了所有的jpeg图片，但是使用 rapid_videocr -i dir 并不能成功运行, 报错如下：

Traceback (most recent call last):
  File "/home/tlss/.pyenv/versions/3.13.1/bin/rapid_videocr", line 8, in <module>
    sys.exit(main())
             ~~~~^^
  File "/home/tlss/.pyenv/versions/3.13.1/lib/python3.13/site-packages/rapid_videocr/rapid_videocr.py", line 320, in main
    extractor(args.img_dir, args.save_dir, args.out_format)
    ~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/tlss/.pyenv/versions/3.13.1/lib/python3.13/site-packages/rapid_videocr/rapid_videocr.py", line 67, in __call__
    srt_result, txt_result = self.single_rec(img_list)
                             ~~~~~~~~~~~~~~~^^^^^^^^^^
  File "/home/tlss/.pyenv/versions/3.13.1/lib/python3.13/site-packages/rapid_videocr/rapid_videocr.py", line 78, in single_rec
    time_str = self.get_time(img_path)
  File "/home/tlss/.pyenv/versions/3.13.1/lib/python3.13/site-packages/rapid_videocr/rapid_videocr.py", line 183, in get_time
    start_str = ':'.join(start_time[:3]) + f',{start_time[3]}'
                                               ~~~~~~~~~~^^^
IndexError: list index out of range

看报错应该是跟文件名相关的，所以想问一下文件名应该如何写？

[SWHL]

的确和文件名有关系。这里输入文件的名字是VideoSubFinder工具的输出。可以参考：
https://github.com/SWHL/RapidVideOCR/tree/main/tests/test_files/RGBImages

[TobisLee]

好的 感谢 能否问一下问什么识别时要跟文件名耦合呢？我看文件名貌似是时间戳+截图区域？这些信息在OCR时会用到吗？