This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
I am afraid the 10x10 resolution is the killer here. Faces are trained on 24x24 pixels, the lowest resolution that contains useful features. But even then we see a lot of models coming out with a larger resolution to capture more detail.
Upscale your image once, to double the size, or until your object reaches at least 24x24 pixels. Then you will see the detector doing something useful.
