AlfredXiangWu / Face_verification_experiment
Original Caffe Version for LightCNN-9. Highly recommend to use PyTorch Version (https://github.com/AlfredXiangWu/LightCNN)
Stars: ✭ 712
Programming Languages
matlab
3953 projects
Labels
Projects that are alternatives of or similar to Face verification experiment
Sphereface
Implementation for <SphereFace: Deep Hypersphere Embedding for Face Recognition> in CVPR'17.
Stars: ✭ 1,483 (+108.29%)
Mutual labels: caffe, face-recognition
Facerecognition
This is an implematation project of face detection and recognition. The face detection using MTCNN algorithm, and recognition using LightenenCNN algorithm.
Stars: ✭ 137 (-80.76%)
Mutual labels: caffe, face-recognition
Seqface
SeqFace : Making full use of sequence information for face recognition
Stars: ✭ 125 (-82.44%)
Mutual labels: caffe, face-recognition
Sphereface Plus
SphereFace+ Implementation for <Learning towards Minimum Hyperspherical Energy> in NIPS'18.
Stars: ✭ 151 (-78.79%)
Mutual labels: caffe, face-recognition
Noiseface
Noise-Tolerant Paradigm for Training Face Recognition CNNs
Stars: ✭ 132 (-81.46%)
Mutual labels: caffe, face-recognition
Largemargin softmax loss
Implementation for <Large-Margin Softmax Loss for Convolutional Neural Networks> in ICML'16.
Stars: ✭ 319 (-55.2%)
Mutual labels: caffe, face-recognition
Pinto model zoo
A repository that shares tuning results of trained models generated by TensorFlow / Keras. Post-training quantization (Weight Quantization, Integer Quantization, Full Integer Quantization, Float16 Quantization), Quantization-aware training. TensorFlow Lite. OpenVINO. CoreML. TensorFlow.js. TF-TRT. MediaPipe. ONNX. [.tflite,.h5,.pb,saved_model,tfjs,tftrt,mlmodel,.xml/.bin, .onnx]
Stars: ✭ 634 (-10.96%)
Mutual labels: caffe
Realtime multi Person pose estimation
Code repo for realtime multi-person pose estimation in CVPR'17 (Oral)
Stars: ✭ 4,760 (+568.54%)
Mutual labels: caffe
Caffe Int8 Convert Tools
Generate a quantization parameter file for ncnn framework int8 inference
Stars: ✭ 483 (-32.16%)
Mutual labels: caffe
Facerecognitiondotnet
The world's simplest facial recognition api for .NET on Windows, MacOS and Linux
Stars: ✭ 672 (-5.62%)
Mutual labels: face-recognition
Test Tube
Python library to easily log experiments and parallelize hyperparameter search for neural networks
Stars: ✭ 663 (-6.88%)
Mutual labels: caffe
Mmdnn
MMdnn is a set of tools to help users inter-operate among different deep learning frameworks. E.g. model conversion and visualization. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML.
Stars: ✭ 5,472 (+668.54%)
Mutual labels: caffe
Enet
ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation
Stars: ✭ 530 (-25.56%)
Mutual labels: caffe
Ffdl
Fabric for Deep Learning (FfDL, pronounced fiddle) is a Deep Learning Platform offering TensorFlow, Caffe, PyTorch etc. as a Service on Kubernetes
Stars: ✭ 640 (-10.11%)
Mutual labels: caffe
Person search
Joint Detection and Identification Feature Learning for Person Search
Stars: ✭ 666 (-6.46%)
Mutual labels: caffe
Mtcnn Caffe
Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Neural Networks
Stars: ✭ 499 (-29.92%)
Mutual labels: caffe
Lightcnn
A Light CNN for Deep Face Representation with Noisy Labels, TIFS 2018
Stars: ✭ 610 (-14.33%)
Mutual labels: face-recognition
A Light CNN for Deep Face Representation with Noisy Labels
Citation
If you use our models, please cite the following paper:
@article{wulight,
title={A Light CNN for Deep Face Representation with Noisy Labels},
author={Wu, Xiang and He, Ran and Sun, Zhenan and Tan, Tieniu}
journal={arXiv preprint arXiv:1511.02683},
year={2015}
}
@article{wu2015lightened,
title={A Lightened CNN for Deep Face Representation},
author={Wu, Xiang and He, Ran and Sun, Zhenan},
journal={arXiv preprint arXiv:1511.02683},
year={2015}
}
@article{wu2015learning,
title={Learning Robust Deep Face Representation},
author={Wu, Xiang},
journal={arXiv preprint arXiv:1507.04844},
year={2015}
}
Updates
- Dec 16, 2016
- The MS-Celeb-1M clean list is uploaded: Baidu Yun, Google Drive.
- Nov 08, 2016
- The prototxt and model C based on caffe-rc3 is updated. The accuracy on LFW achieves 98.80% and the [email protected]=0 obtains 94.97%.
- The performance of set 1 on MegaFace achieves 65.532% for rank-1 accuracy and 75.854% for [email protected]=10^-6.
- Nov 26, 2015
- The prototxt and model B is updated and the accuracy on LFW achieves 98.13% for a single net without training on LFW.
- Aug 13, 2015
- Evaluation of LFW for identification protocols is published.
- Jun 11, 2015
- The prototxt and model A is released. The accuracy on LFW achieves 97.77%.
Overview
The Deep Face Representation Experiment is based on Convolution Neural Network to learn a robust feature for face verification task. The popular deep learning framework caffe is used for training on face datasets such as CASIA-WebFace, VGG-Face and MS-Celeb-1M. And the feature extraction is realized by python code caffe_ftr.py.
Structure
- Code
- data pre-processing and evaluation code
- Model
- caffemodel.
- The model A and B is trained on CASIA-WebFace by caffe-rc.
- The model C is trained on MS-Celeb-1M by caffe-rc3.
- caffemodel.
- Proto
- Lightened CNN implementations by caffe
- Results
- LFW features
Description
Data Pre-processing
- Download face dataset such as CASIA-WebFace, VGG-Face and MS-Celeb-1M.
- All face images are converted to gray-scale images and normalized to 144x144 according to landmarks.
- According to the 5 facial points, we not only rotate two eye points horizontally but also set the distance between the midpoint of eyes and the midpoint of mouth(ec_mc_y), and the y axis of midpoint of eyes(ec_y) .
Dataset | size | ec_mc_y | ec_y |
---|---|---|---|
Training set | 144x144 | 48 | 48 |
Testing set | 128x128 | 48 | 40 |
Training
- The model is trained by open source deep learning framework caffe.
- The network configuration is showed in "proto" file and the trained model is showed in "model" file.
Evaluation
- The model is evaluated on LFW which is a popular data set for face verification task.
- The extracted features and lfw testing pairs are located in "results" file.
- To evaluate the model, the matlab code or other ROC evaluation code can be used.
- The model is also evaluated on MegaFace. The dataset and evaluation code can be downloaded from http://megaface.cs.washington.edu/
Results
The single convolution net testing is evaluated on unsupervised setting only computing cosine similarity for lfw pairs.
Model | 100% - EER | [email protected]=1% | [email protected]=0.1% | [email protected]=0 | Rank-1 | [email protected]=1% |
---|---|---|---|---|---|---|
A | 97.77% | 94.80% | 84.37% | 43.17% | 84.79% | 63.09% |
B | 98.13% | 96.73% | 87.13% | 64.33% | 89.21% | 69.46% |
C | 98.80% | 98.60% | 96.77% | 94.97% | 93.80% | 84.40% |
The details are published as a technical report on arXiv.
The released models are only allowed for non-commercial use.
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].