Face Detection#

Face detection is the task of detecting faces in a photo or video (and distinguishing them from other objects). We provide the kornia.contrib.FaceDetector to perform multi-face detection in real-time using the YuNet model.

Learn more: https://paperswithcode.com/task/face-detection

Using our API you can easily detect faces in images as shown below:

import torch
import kornia as K
from kornia.contrib import FaceDetector, FaceDetectorResult
from kornia.io import ImageLoadType

# select the device
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

# load the image as a (1, 3, H, W) RGB float tensor in [0, 255], the range FaceDetector expects;
# it flips the channels to the BGR order YuNet was trained on
img = K.io.load_image("image.jpg", ImageLoadType.RGB32, device=device)[None] * 255.0

# create the detector and find the faces
face_detection = FaceDetector().to(device)

with torch.no_grad():
    dets = face_detection(img)
dets = [FaceDetectorResult(o) for o in dets[0]]

for det in dets:
    print(det.score, det.top_left, det.bottom_right)

Play with the detector yourself and generate new images with this tutorial.