olojuwin / Facerecognize For Mobile Phone
适用于移动端的人脸识别模型,计算量与mobilefacenet相同,但megaface上提升了2%+
Stars: ✭ 229
Projects that are alternatives of or similar to Facerecognize For Mobile Phone
Face-Mask-Detection-PyTorch
A real-time face mask detector based on computer vision and deep learning, created using Pytorch and OpenCV
Stars: ✭ 12 (-94.76%)
Mutual labels: face-recognition, mobilenetv2
Ownphotos
Self hosted alternative to Google Photos
Stars: ✭ 2,587 (+1029.69%)
Mutual labels: face-recognition
Face.evolve.pytorch
🔥🔥High-Performance Face Recognition Library on PaddlePaddle & PyTorch🔥🔥
Stars: ✭ 2,719 (+1087.34%)
Mutual labels: face-recognition
Marvel
Marvel - Face Recognition With Android & OpenCV
Stars: ✭ 199 (-13.1%)
Mutual labels: face-recognition
Howdy
🛡️ Windows Hello™ style facial authentication for Linux
Stars: ✭ 3,237 (+1313.54%)
Mutual labels: face-recognition
C Ms Celeb
A clean version (wash list) of MS-Celeb-1M face dataset, containing 6,464,018 face images of 94,682 celebrities
Stars: ✭ 227 (-0.87%)
Mutual labels: face-recognition
Maskinsightface
基于人脸关键区域提取的人脸识别(LFW:99.82%+ CFP_FP:98.50%+ AgeDB30:98.25%+)
Stars: ✭ 221 (-3.49%)
Mutual labels: face-recognition
Mobilenet V2
A Complete and Simple Implementation of MobileNet-V2 in PyTorch
Stars: ✭ 206 (-10.04%)
Mutual labels: mobilenetv2
Esp32 Cam Webserver
Expanded version of the Espressif ESP webcam
Stars: ✭ 200 (-12.66%)
Mutual labels: face-recognition
Arcface Multiplex Recognition
适用于复杂场景的人脸识别身份认证系统
Stars: ✭ 200 (-12.66%)
Mutual labels: face-recognition
Lightnetplusplus
LightNet++: Boosted Light-weighted Networks for Real-time Semantic Segmentation
Stars: ✭ 218 (-4.8%)
Mutual labels: mobilenetv2
Mediadevices
Go implementation of the MediaDevices API.
Stars: ✭ 197 (-13.97%)
Mutual labels: face-recognition
Faceimagequality
Code and information for face image quality assessment with SER-FIQ
Stars: ✭ 223 (-2.62%)
Mutual labels: face-recognition
Pytorch Deeplab Xception
DeepLab v3+ model in PyTorch. Support different backbones.
Stars: ✭ 2,466 (+976.86%)
Mutual labels: mobilenetv2
Facerecognition
Implement face recognition using PCA, LDA and LPP
Stars: ✭ 206 (-10.04%)
Mutual labels: face-recognition
Mobilefacenet pytorch
MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices
Stars: ✭ 209 (-8.73%)
Mutual labels: face-recognition
Openface
Face recognition with deep neural networks.
Stars: ✭ 14,249 (+6122.27%)
Mutual labels: face-recognition
Insightface Tensorflow
Tensoflow implementation of InsightFace (ArcFace: Additive Angular Margin Loss for Deep Face Recognition).
Stars: ✭ 228 (-0.44%)
Mutual labels: face-recognition
facerecognize-for-mobile-phone
适用于移动端的人脸识别模型,计算量于mobilefacenet相同,但megaface上提升了2%+。我训练的mobilefacenet在megaface VER上比原作者 高出了2%,因为训练的数据和方法不一样。 2020.02.24新增GPU模型r100fc,megaface rank1 99.00%+
模型在各个数据集上表现如下:
Methods | Flops (112x112) | LFW | CFP-FP | AgeDB | Megaface-Id | [email protected] | 备 注 |
---|---|---|---|---|---|---|---|
MobileFaceNet440,R | 440M | 99.70+ | 96.70+ | 96.95+ | 92.85+ | 94.20+ | 未开源 |
ZW350 | 356M | 99.70+ | 96.82+ | 97.00+ | 93.90+ | 94.70+ | 未开源 |
ZW400 | 404M | 99.70+ | 96.95+ | 97.00+ | 94.46+ | 95.60+ | 未开源 |
MobileFaceNet600,R | 612M | 99.76+ | 97.60+ | 97.50+ | 95.14+ | 95.98+ | 已开源 |
ZW440 | 444M | 99.76+ | 97.30+ | 97.40+ | 95.25+ | 96.00+ | 已开源 |
r100fc | 24G | 99.86+ | 99.10+ | 98.50+ | 99.00+ | 98.80+ | 已经取消开源 |
Megaface测试结果图
zw440-ver
速度比对测试
设备:i5-6500
Methods | Openvino | opencv单线程 |
---|---|---|
MobileFaceNet600,R | 6ms | 141ms |
ZW440 | 7ms | 80ms |
移动设备
经过测试,zw440并没有Mobilefacenet600M快.感谢moli的测试
模型地址
模型包含mxnet ncnn caffe 三种格式 Baidu Drive 提取码:b0dm
r100 Baidu Drive 提取码:224u
训练数据
https://github.com/deepinsight/insightface/tree/master/iccv19-challenge
参考项目
https://github.com/deepinsight/insightface
https://github.com/happynear/FaceVerification
https://github.com/Tencent/ncnn
https://github.com/cypw/MXNet2Caffe
Todo
没有做速度方面考虑,后期跟进改善。
License
BSD 3 Clause
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].