I am teaching a course on Computer Vision for Graduate Management Students using KNIME Image processing extension.
Can I get a work flow how to train and classify Images containg more than one object say a person in front a car in the background and some trees in other part all in one IMAGE.
Hi @Chandra_S,
that is quite the challenging classification task, the quickest solution would be using something like:
https://hub.knime.com/knime/spaces/Examples/latest/04_Analytics/14_Deep_Learning/02_Keras/01_Classify_images_using_InceptionV3 and using the top k classes instead of only the best prediction.
best,
Gabriel
Dear @Chandra_S,
We ran a webinar about semantic image segmentation a few months ago. You can re-watch the recording at https://youtu.be/_Fmkqwrt_Gk?t=2684 and download the workflow that Benjamin is describing from https://kni.me/w/-Bs3paHwNKVHah3Q (just copy this link via Ctrl+C and go into your KNIME and paste into one of your folders in the KNIME Explorer).
Best regards,
Stefan
