Is KNIME a platform that can or is being used for mining of video content and in particular classification of patterns within the video through supervised machine learning ? I watched the image processing video and it got me thinking if it was possible to do something similar with video
Would the process be to break the video into frames and process it as separate but connected images?
Hi mobmsc,
the KNIME Image Processing Extension [1] can actually also process videos, without breaking the data into individual pieces. The car-counting example [2] contains both, a a video broken into images and later we turn the individual images into a video. There are actually many more examples with videos, 3d-data etc.
If you like, you can explain a bit more in detail what you want to do (and even post some example images) and maybe I can help you getting started.
Cheers,
Christian
[1] https://tech.knime.org/community/image-processing
[2] https://tech.knime.org/community/image-processing#exampleworkflows
Its only a germ of an idea at the moment for a project but its to provide a dataset of short videos already classified and try to mine a longer video to identify repeated instances of that pattern.
I don't have examples but I'll take a look at the examples you suggested (after increasing the java heap size by editing the knime.ini file as per https://tech.knime.org/faq#q4_2)
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.