Will decreasing the video dimension will reduce the resolution? [closed]
In order to speed up face detection process, some will reduce the video or image dimension. Will it makes the resolution become worst and effect the face recognition process?
the face cascades were trained on 24x24 pixel images, so that is the minimal size you can detect in your video.
So complementing @berak's comment, the answer is "it depends". Suppose you have an original 800x600 video, where the smallest possible face is of size 80x80. If you re-size the video to 400x300, the smallest face size will become 40x40. As face cascades detect 24x24 faces, there is no problem, and you will speed up execution. If you resize the video to 200x150, minimum face size is 20x20, and those smallest faces won't be detected