detect_image is detecting multiple faces in an image with a single face

[awymanquant]

I just started using Py-Feat to analyze participants emotions while taking a stress task. However, for some participants, the algorithm is detecting multiple faces in the image even though it's only one person in front of the camera. From plot_detections, it looks like the algorithm is dividing the person's face into two different faces (top half of face and bottom half of face, with the nose region overlapping in both).

I apologize if this issue has been solved already, but is there a way to coerce the algorithm to only detect one face instead of searching for all possible faces?

Thanks so much!

[ljchang]

Hi @awymanquant , there are a few different things to try to reduce this undesirable behavior.

1. Adjust face_detection_threshold=0.5
2. Try a different face_model in the Detector()
3. We have also found that there are more false positives in high res images. you could also try downsampling the image before running detection.