facebookresearch
Text Prompt "person/human" Not Working for Segmentation (Works with "shoe")
In examples/sam3_image_predictor_example.ipynb, the text prompt segmentation works correctly with "shoe", but no valid mask is generated when using "person" or "human" (with images containing clear human subjects).
Could you provide guidance on how to properly use text prompts for person/human segmentation? Thanks!
After testing, the text prompt "a person" now works for human segmentation (where "person" or "human" alone failed earlier)!
It's quite hard to tell what's happening without looking at the image. One thing you can play with is the detection threshold. It's 0.5 by default, you can try lowering to 0.3 to improve recall.
same for me, even 'shoes' is not working, but 'shoe' works well. Seems very senstive to text
