Am I using the correct approach to find faces in a photo?

Question

Machine learning

There is a task to find a face in a photo, as I designed the solution: I
calculate a descriptor for the hog image, then I train a binary svm classifier to identify whether the hog image descriptor is a face, offering the classifier 500 hog descriptors of images with a face, and 100 images each lantern, cars and other unnecessary content
Then I run a 64 * 128 window over the 128 * 256 hog descriptor with a step of 4 pixels in width and height, as a result, the window appears on the image 512 times, and constantly determines how similar the captured part is to the template hog- face descriptor, and then gives out the cut block, on which the similarity index was maximum, thereby cutting out the face
. Actually the question is:
Am I doing the right thing and are there any better ways?

Answer 1

https://www.pyimagesearch.com/2018/09/24/opencv-fa...