You are viewing limited content. For full access, please sign in.

Question

Question

Is there a way to automate metadata for current documents?

Laserfiche Version 10 Self-Hosted How To
Updated February 19
asked on February 19

Many years of transcripts were transferred from our network drive to LF in pdf form.  No metadata was ever filled out.  I am in the process of generating OCR and pages and deleting the pdf.  Metadata is being applied to new transcripts so that is not a problem.  Is there a way to apply metadata to the transcripts already in LF without needing to manually go into each file to fill out the metadat?

0 0

Replies

replied on February 19

Hi Shari,

If you have Quick Fields, you can use the Laserfiche Capture Engine to point Quick Fields to the repository and find documents there to "scan." If the metadata you want to capture off these images is clear and readable and reliably in the same spot every time, you can set up a Quick Fields session to find those pieces of information and fill out field values for you.

Even if the data isn't in the same spot every time, if the text surrounding it is consistent, you could use the document text that you're getting from OCR to perform some pattern matching to try to extract what you want. 

So while I think the answer to your question is "yes," one method might be better than another depending on what these images look like. Does that help at all?

2 0
View 2 previous replies
replied on February 19

I will need to get access to Quick Fields first.  Will it work on items already in the repository?  We have several years of handwritten transcripts and also have several different formats of transcripts.  These could be grouped together in different sessions.

 

Thank you for your reply!

0 0
replied on February 19 • Show version history

Handwritten text will be a different beast if that text is something you want to extract and put into a field value. There might be a way to do it, but OCR will only consistently work on typed text. Laserfiche Cloud has some built-in functionality that works surprisingly well with handwritten text, but I don't believe that same functionality is available in self-hosted systems.

But yes, with Quick Fields, you set a scan source, which can be a network folder that Quick Fields pulls files from, a scanner hooked up to your PC, or you can point it to your repository and process documents that you already have in Laserfiche. 

1 0
replied on February 19

That is great information!

0 0
replied on February 19

Something else you could do is look at the document text that has already been OCR'd and see if the text you need is there. For example, if someone's name is what you want to capture, the document text might have "NAME: SMITH, JOHN" and you could use the Workflow application to look for that and capture what comes after "NAME:." Unlikely it's that simple, but could be a place to start as that wouldn't need to necessarily involve Quick Fields.

All just depends on what these things look like. 

2 0
replied on February 19

Thank you so much!

0 0
You are not allowed to follow up in this post.

Sign in to reply to this post.

Details
Related Posts
Loading ...