How can we tell Object is Icon? TextBox, etc?

#12
by Verfinux - opened

Is there an Object Type return, where can I get this, API ? like object is a Text box that we can enter text, Icon that can click?
if is a Windows standard Close button, which windows title does it belongs to, so we will not close the wrong windows?

Microsoft org

Hi @Verfinux ,there is an object type returned by the model. Feel free to try out our demo: https://huggingface.co/spaces/microsoft/OmniParser. It also output the bbox of each detected elements

Sign up or log in to comment