Face landmark detection (Userscript Example)

Mode Homepage



-> Model generally assumes the face will fill a significant part of the screen
-> Note this should actually create a keypoint type instance. However you must have a keypoint instance with a similar number of points defined in your project.


const model = await faceLandmarksDetection.load();
let canvas = diffgram.get_new_canvas()
const predictions = await model.estimateFaces({
  input: canvas
for (let i = 0; i < predictions.length; i++) {
      const keypoints = predictions[i].scaledMesh;

  // Log facial keypoints.
  for (let i = 0; i < keypoints.length; i++) {
    const [x, y, z] = keypoints[i];

    //console.log(`Keypoint ${i}: [${x}, ${y}, ${z}]`);
    diffgram.create_instance_from_keypoints(x, y)

Did this page help you?