这是一个利用facenet和retinaface实现人脸识别的库，可以进行在线的人脸识别。

Last update: Dec 30, 2022

Overview

Facenet+Retinaface：人脸识别模型在Pytorch当中的实现

注意事项

该库中包含了两个网络，分别是retinaface和facenet。二者使用不同的权值。在使用网络时一定要注意权值的选择，以及主干与权值的匹配。

所需环境

pytorch==1.2.0

文件下载

预测所需的权值文件可以在百度云下载。
链接: https://pan.baidu.com/s/1iTo4_x0DHg0GoTUQWduMZw 提取码: dmw6

预测步骤

本项目自带主干为mobilenet的retinaface模型与facenet模型。可以直接运行，如果想要使用主干为resnet50的retinafa和主干为inception_resnetv1的facenet模型需要。
在retinaface.py文件里面，在如下部分修改model_path和backbone使其对应训练好的文件。

_defaults = {
    "retinaface_model_path" : 'model_data/Retinaface_mobilenet0.25.pth',
    #-----------------------------------#
    #   可选retinaface_backbone有
    #   mobilenet和resnet50
    #-----------------------------------#
    "retinaface_backbone"   : "mobilenet",
    "confidence"            : 0.5,
    "iou"                   : 0.3,
    #----------------------------------------------------------------------#
    #   是否需要进行图像大小限制。
    #   输入图像大小会大幅度地影响FPS，想加快检测速度可以减少input_shape。
    #   开启后，会将输入图像的大小限制为input_shape。否则使用原图进行预测。
    #   会导致检测结果偏差，主干为resnet50不存在此问题。
    #   可根据输入图像的大小自行调整input_shape，注意为32的倍数，如[640, 640, 3]
    #----------------------------------------------------------------------#
    "retinaface_input_shape": [640, 640, 3],
    #-----------------------------------#
    #   是否需要进行图像大小限制
    #-----------------------------------#
    "letterbox_image"       : True,
    
    "facenet_model_path"    : 'facenet_inception_resnetv1.pth',
    #-----------------------------------#
    #   可选facenet_backbone有
    #   mobilenet和inception_resnetv1
    #-----------------------------------#
    "facenet_backbone"      : "inception_resnetv1",
    "facenet_input_shape"   : [160,160,3],
    "facenet_threhold"      : 0.9,

    "cuda"                  : True
}

运行encoding.py，对face_dataset里面的图片进行编码，face_dataset的命名规则为XXX_1.jpg、XXX_2.jpg。最终在model_data文件夹下生成对应的数据库人脸编码数据文件。
运行predict.py，输入下述文字，可直接预测。

img/zhangxueyou.jpg

利用video.py可进行摄像头检测。

Reference

https://github.com/biubug6/Pytorch_Retinaface

这是一个利用facenet和retinaface实现人脸识别的库，可以进行在线的人脸识别。

Related tags

Overview

Facenet+Retinaface：人脸识别模型在Pytorch当中的实现

目录

注意事项

所需环境

文件下载

预测步骤

Reference

Owner

Bubbliiiing

MoveNet Single Pose on DepthAI

ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation.

An official PyTorch implementation of the TKDE paper "Self-Supervised Graph Representation Learning via Topology Transformations".

We propose a new method for effective shadow removal by regarding it as an exposure fusion problem.

Bald-to-Hairy Translation Using CycleGAN

Segmentation-Aware Convolutional Networks Using Local Attention Masks

The dataset and source code for our paper: "Did You Ask a Good Question? A Cross-Domain Question IntentionClassification Benchmark for Text-to-SQL"

Ensembling Off-the-shelf Models for GAN Training

Homepage of paper: Paint Transformer: Feed Forward Neural Painting with Stroke Prediction, ICCV 2021.

Pervasive Attention: 2D Convolutional Networks for Sequence-to-Sequence Prediction

Awesome Long-Tailed Learning

Code for our TKDE paper "Understanding WeChat User Preferences and “Wow” Diffusion"

PyTorch implementation of SimSiam: Exploring Simple Siamese Representation Learning

[ICLR2021oral] Rethinking Architecture Selection in Differentiable NAS

Cweqgen - The CW Equation Generator

Self-supervised Multi-modal Hybrid Fusion Network for Brain Tumor Segmentation

Code for "Intra-hour Photovoltaic Generation Forecasting based on Multi-source Data and Deep Learning Methods."

a pytorch implementation of auto-punctuation learned character by character

Pytorch Implementation of "Diagonal Attention and Style-based GAN for Content-Style disentanglement in image generation and translation" (ICCV 2021)

Official code of paper: MovingFashion: a Benchmark for the Video-to-Shop Challenge