Open dahlheim2 opened 2 years ago
Errr - no - the output should be like
Where you can see the payload is an array of objects - each with its own bounding box.
well, yes, thank you for the attention and response and for the careful explanation but i'm aware of that and a direct output of msg.payload from the node seems to indicate more than one person detected often in msg.classes but only one bounding box per image is output in msg.payload.
please can you run the example supplied with the node and show me the debug output expanded as per above.
thanks once again, for pointing out that example, which i had not yet stumbled across, having installed the node from the node-red palette gui feature. it helped me sort out my own confusion.
i guess i just wonder if this makes sense?
indeed that does not look correct... but maybe the confidence score for one of the persons was below the confidence limit... (If so it should not have been counted in the classes - but may be a reason for why it's not in the payload).
what a great node, thanks!
i notice when more than one object has been detected in an image, the payload contains the information it should as described by the documentation. but if more than one object is detected, for example more than one person, only one bounding box's coordinates are described in the resultant msg.payload object. perhaps i am missing something?