I was trying to understand if Knime has some useful instruments for extract information from images/videos.
Maybe using some image recognition algorithms or something like that.
If there is something, can someone share any link/examples/resources about it?
Since from the few knime examples I have seen only numbers as image recognition with convolutional neural networks and DL4J deep learning extension. Is it possible to process and recognize something different from simple handwritten numbers?
I mean something at a higher level, like faces, cars, birds, cats, dogs, persons, places etc..?
there are some example workflows for the new Keras integration that was introduced with KNIME 3.5. You can find them here at the bottom of the page. Especially the two "Classify images using..." workflows should be interesting.
As far as I know, updating KNIME from 3.4 to 3.5 is currently disabled on Mac. You''ll have to reinstall it from scratch, unfortunately. Please note that support for Linux 32 bit has been discontinued in case you run Ubuntu 32 bit. For Ubuntu 64 bit, updating to 3.5 should work normally via "File -> Update KNIME...". If it doesn't, please check if the listed software sites under "File -> Preferences -> Install/Update -> Available Software Sites" contain a "3.5" in their "Location" column and are enabled.
Regarding the mac version I have just reinstalled it from scratch. Regarding the ubuntu version (64 bit) I cannot find the 3.5 version under "File -> Preferences -> Install/Update -> Available Software Sites". I have attached a screenshot.