insightface | State-of-the-art 2D and 3D Face Analysis Project | Computer Vision library

faces_emore/

​    train.idx

​    train.rec

​    property

​    lfw.bin

​    cfp_fp.bin

​    agedb_30.bin

python tools/mx_recordio_2_ofrecord_shuffled_npart.py  --data_dir datasets/faces_emore --output_filepath faces_emore/ofrecord/train --part_  

models
    ├── dlib-19.17
    │   ├── models
    │       ├── dlib_face_recognition_resnet_model_v1.dat
    │       └── shape_predictor_5_face_landmarks.dat
    ├── facenet
    │   ├── models  

pip install retina-face

from retinaface import RetinaFace
resp = RetinaFace.detect_faces("img1.jpg")

{
    "face_1": {
        "score": 0.9993440508842468,
        "facial_area": [155, 81, 434, 443],
        "landmarks": {
          "right_eye": [2  

import sys
import os
import argparse
import onnx
import json
import mxnet as mx
from onnx import helper
from onnx import TensorProto
from onnx import numpy_helper
import onnxruntime
import cv2

print('mxnet version:', mx.__version__)
print('onnx vers  

import datetime
import numpy as np
import os
import os.path as osp
import glob
import cv2
import insightface

assert insightface.__version__>='0.4'


def detect_person(img, detector):
    bboxes, kpss = detector.detect(img)
    bboxes = np.round(b  

import datetime
import numpy as np
import os
import os.path as osp
import glob
import cv2
import insightface
from insightface.app import FaceAnalysis
from insightface.data import get_image as ins_get_image


assert insightface.__version__>='0.7'

  

#le = LabelEncoder()
#labels = le.fit_transform(data["names"])
#num_classes = len(np.unique(labels))
#labels = labels.reshape(-1, 1)
#one_hot_encoder = OneHotEncoder(categorical_features = [0])
#labels = one_hot_encoder.fit_transform(label

face_imgs_resized = np.ones((128,224,224,3))
face_imgs_resized = np.array(face_imgs_resized)
face_imgs_resized = np.rollaxis(face_imgs_resized, 3, 1)
data = mx.nd.array(face_imgs_resized)
db = mx.io.DataBatch(data=mx.nd.split_v2(data, 4))