Hi. An updated OpenCV package has been released, which includes an example of the FaceDetector Safe Model. A very cool solution, the stability of facial recognition is an order of magnitude higher. I also liked the presence of the Faces pin in node Detect Faces. You can view the result of the detected face. But there is one problem: the image resolution is constantly changing. Is there any way to fix the resolution of the output image?
any chance you can demo this in a patch?
Perhaps I asked the question incorrectly. I’m still building my facial recognition software. And I need the resolution of the image to match the resolution of the images used to train the model.
please chekc your video. it onyl shows a single frame for me. what is it supposed to show?
Something really doesn’t show what is needed in the video. Maybe an excerpt of the patch will help to understand my problem
Face Detector Caffe Model.vl (22.5 KB)
What doesn’t look right here is the Join [OpenCV.Utils] node. What are you expecting it to do?
I want to get an image of a face to identify it.
hovering the Join node, tells you what it’s doing:
this doesn’t sound like what you want here. instead i believe you want to get (for a start at least) individual images out of the spread of detected images. for this can you use e.g. a GetSlice node…
Thanks for the help, now I will read the description more carefully. And another question: is there any way to recognize a blurred image using opencv?