This documentation seems to be more geared towards mobile app developers but if this service is available to you it might be worth looking at https://developers.google.com/ml-kit/vision/selfie-segmentation - though it might still crop a silhouette… either way sharing because I was looking at it
it is mobile use case but react-native, so this probably won’t work since it is a native OS solution
yeah i got there from a deprecated firebase thing so I was hoping it wouldn’t be a big jump
maybe it was easier when things weren’t so good at detecting faces … you could just be like “eh, can’t find anything.” … vs now they’re like “I think I see a forhead”
OK So we want the GCP equivalent to https://docs.aws.amazon.com/rekognition/latest/APIReference/API_FaceOccluded.html
I wonder about skipping the faces part and looking at objects/labels
i looked at that a bit but the labels are like so many…
I did get “person” on this one
i did on the other one too I just didn’t see it
yea person does not always show
i just dont know how these labels work. its just so random
I was hoping Face would be in the label or something
yea exactly
give me face, eyes, nose, mouth
but its hit and miss
sometimes it has none of those, no facial parts at all
Well… I couldn’t really help you but at least I learned how to make sure AI thinks my linkedin profile picture is MVP :rolling_on_the_floor_laughing:
haha thank you …
learn something new every day!
