4 ms·
Extracting Structured Data from Templatic Documents
- xnx 6y agoGood to remember blog posts like this for all those who claim Google isn't innovating or investing in search. This is the type of infrastructure that goes in to extracting useful information from the web.
- vitorbaptistaa 6y agoThe trained model doesn't seem to be available? Does anyone know if this or a similar model is available somewhere?
- AndrewKemendo 6y agoYears ago I tried hacking something together like this - primarily to read labels on packaged food products - but the primary roadblock I hit was with the FOSS OCR solutions not being anywhere near good enough to be reliable. Mind you this was a few years ago and I was primarily testing with pytesseract. I would be curious if this team actually used the Google OCR API or an internally tuned one that isn't GA, and how that differs FOSS Tesseract.
- The_rationalist 6y agoThe state of the art is from 2017 https://paperswithcode.com/sota/optical-character-recognition-on-fsns-test https://paperswithcode.com/sota/optical-character-recognitio...
- combatentropy 6y agoThis reminds me of some of Sergey Brin's research papers (https://scholar.google.com/scholar?q=sergey+brin https://scholar.google.com/scholar?q=sergey+brin), like "What Can You Do with the Web in Your Pocket?" (https://www.cs.princeton.edu/courses/archive/fall02/cs597A/papers/brin98what.pdf https://www.cs.princeton.edu/courses/archive/fall02/cs597A/p...)
- mukuz 6y agoGoogle will be able to provide us with a lot more information extracted from our mails. Like it currently does for delivery tracking and reservations.