Has the extraction of text-related feature been implement?

[Liuhaai]

It seems that the extractor "OCRSearch" and "DescriptionTextSearch" don't work when I do extract operation on video files.
Here are the errors :

2021-03-16 12:36:04.791 [pool-4-thread-4] FATAL o.v.c.s.r.ExtractionTask - EXTRACTION ERROR in OCRSearch: for segmentID v_0000003_2
java.lang.UnsupportedOperationException: Not supported by default
at org.vitrivr.cineast.core.features.abstracts.AbstractTextRetriever.processSegment(AbstractTextRetriever.java:85)
at org.vitrivr.cineast.standalone.runtime.ExtractionTask.run(ExtractionTask.java:28)

The error output is come from this code in AbstractTextRetriever.java:

  public void processSegment(SegmentContainer shot) {
    throw new UnsupportedOperationException("Not supported by default");
  }

[lucaro]

No, neither of these is currently implemented, they rely on external data. In the case of the description search, this can't be implemented as an extractor, since it targets descriptions which may accompany a video file, such as what you find on youtube or vimeo, rather than being part of the video itself.