Best way to do Table Detection in 2024 with TF-ID - Quick Tutorial with Code

[@7890shawn]

Hey William, really a great video. This really helped me in better table detection for my Document Layout Analysis project which is the first step for the same. Your video helped me understand its potential for Structure Recognition part as well. Florence 2 really is an interesting model. Keep these videos coming, definitely earned my Sub. 💯

[@python-programming]

I'm so happy to hear that! Thanks for the comment. If you ever want to see anything in particular on the channel, please don't hesitate to let me know!

[@7890shawn]

@@python-programming Hey William, thanks for this. Since you’ve asked and I’m greedy for more knowledge 😝 I want to know your thoughts on Structure Recognition after tables are detected. Have the layout information preserved while extracting tables. Would love a series on this as well. I’m sure there are folks like me who would be interested 😎🥂

[@python-programming]

This is great to know! Thanks! As luck would have it, I'm in the middle of preparing such a series. Trying to decide on the best stack for the video series.

[@7890shawn]

@@python-programming looking forward to it already 🙌🏻😎

[@VenkatesanVenkat-fd4hg]

Great share....

[@python-programming]

Thanks! Glad you like it!

[@philtoa334]

Very Nice .

[@flyingzeppo]

Since you asked for suggestions for future videos, I have 3 ideas. First, since knowing grammar is important for NLP work, maybe a video on intermediate to advance English grammar. Not just, what a noun or verb is, but ideas such as coordinating clauses, tenses, antecedents, etc. My second idea would be a video on using grammar and NLP to pull out the participants of a certain verb in a large body of text. Such as pulling out the people's names who speak to each other when the sentences could be "John speaks to Mary." or "John and Mary are speaking." or "Mary, whom John speaks to every week, said...." etc. And my third suggestion would be a video on Latin grammar. Whatever ideas for videos you have I'll watch 'em. This is one of my favorite KZbin channels. Keep it up. Thanks.

[@python-programming]

I really really like all three of these ideas!! I'll start thinking about how to structure them. I really like the idea of a follow-up video on LatinCy which would be important for the Latin video. I think there's a new update on the models coming soon!