Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/peteryuX/arcface-tf2
ArcFace unofficial Implemented in Tensorflow 2.0+ (ResNet50, MobileNetV2). "ArcFace: Additive Angular Margin Loss for Deep Face Recognition" Published in CVPR 2019. With Colab.
https://github.com/peteryuX/arcface-tf2
arcface arcface-tf2 colab colab-notebook deep-face-recognition face-recognition tensorflow tf2
Last synced: about 2 months ago
JSON representation
ArcFace unofficial Implemented in Tensorflow 2.0+ (ResNet50, MobileNetV2). "ArcFace: Additive Angular Margin Loss for Deep Face Recognition" Published in CVPR 2019. With Colab.
- Host: GitHub
- URL: https://github.com/peteryuX/arcface-tf2
- Owner: peteryuX
- License: mit
- Created: 2019-08-24T03:11:22.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2022-11-21T21:17:05.000Z (about 2 years ago)
- Last Synced: 2024-08-01T03:17:02.207Z (4 months ago)
- Topics: arcface, arcface-tf2, colab, colab-notebook, deep-face-recognition, face-recognition, tensorflow, tf2
- Language: Python
- Homepage:
- Size: 112 KB
- Stars: 258
- Watchers: 11
- Forks: 60
- Open Issues: 22
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-tensorflow-2 - ArcFace: Additive Angular Margin Loss for Deep Face Recognition
README
# [arcface-tf2](https://github.com/peteryuX/arcface-tf2)
[](https://lgtm.com/projects/g/peteryuX/arcface-tf2/context:python)
[](https://colab.research.google.com/github/peteryuX/arcface-tf2/blob/master/notebooks/colab-github-demo.ipynb)
:fire: ArcFace (Additive Angular Margin Loss for Deep Face Recognition, published in CVPR 2019) implemented in Tensorflow 2.0+. This is an unofficial implementation. :fire:
> Additive Angular Margin Loss(ArcFace) has a clear geometric interpretation due to the exact correspondence to the geodesic distance on the hypersphere, and consistently outperforms the state-of-the-art and can be easily implemented with negligible computational overhead.
Original Paper: [Arxiv](https://arxiv.org/abs/1801.07698) [CVPR2019](http://openaccess.thecvf.com/content_CVPR_2019/html/Deng_ArcFace_Additive_Angular_Margin_Loss_for_Deep_Face_Recognition_CVPR_2019_paper.html)
Offical Implementation: [MXNet](https://github.com/deepinsight/insightface)
****
## Contents
:bookmark_tabs:
* [Installation](#Installation)
* [Data Preparing](#Data-Preparing)
* [Training and Testing](#Training-and-Testing)
* [Benchmark and Models](#Benchmark-and-Models)
* [References](#References)
## Installation
:pizza:
Create a new python virtual environment by [Anaconda](https://www.anaconda.com/) or just use pip in your python environment and then clone this repository as following.
### Clone this repo
```bash
git clone https://github.com/peteryuX/arcface-tf2.git
cd arcface-tf2
```
### Conda
```bash
conda env create -f environment.yml
conda activate arcface-tf2
```
### Pip
```bash
pip install -r requirements.txt
```
****
## Data Preparing
:beer:
All datasets used in this repository can be found from [face.evoLVe.PyTorch's Data-Zoo](https://github.com/ZhaoJ9014/face.evoLVe.PyTorch#Data-Zoo).
Note:
- Both training and testing dataset are "Align_112x112" version.
### Training Dataset
Download [MS-Celeb-1M](https://drive.google.com/file/d/1X202mvYe5tiXFhOx82z4rPiPogXD435i/view?usp=sharing) datasets, then extract and convert them to tfrecord as traning data as following.
```bash
# Binary Image: convert really slow, but loading faster when traning.
python data/convert_train_binary_tfrecord.py --dataset_path="/path/to/ms1m_align_112/imgs" --output_path="./data/ms1m_bin.tfrecord"
# Online Image Loading: convert really fast, but loading slower when training.
python data/convert_train_tfrecord.py --dataset_path="/path/to/ms1m_align_112/imgs" --output_path="./data/ms1m.tfrecord"
```
Note:
- You can run `python ./dataset_checker.py` to check if the dataloader work.
### Testing Dataset
Download [LFW](https://drive.google.com/file/d/1WO5Meh_yAau00Gm2Rz2Pc0SRldLQYigT/view?usp=sharing), [Aged30](https://drive.google.com/file/d/1AoZrZfym5ZhdTyKSxD0qxa7Xrp2Q1ftp/view?usp=sharing) and [CFP-FP](https://drive.google.com/file/d/1-sDn79lTegXRNhFuRnIRsgdU88cBfW6V/view?usp=sharing) datasets, then extract them to `/your/path/to/test_dataset`. These testing data are already binary files, so it's not necessary to do any preprocessing. The directory structure should be like bellow.
```
/your/path/to/test_dataset/
-> lfw_align_112/lfw
-> data/
-> meta/
-> ...
-> agedb_align_112/agedb_30
-> ...
-> cfp_align_112/cfp_fp
-> ...
```
****
## Training and Testing
:lollipop:
You can modify your own dataset path or other settings of model in [./configs/*.yaml](https://github.com/peteryuX/arcface-tf2/tree/master/configs) for training and testing, which like below.
```python
# general (shared both in training and testing)
batch_size: 128
input_size: 112
embd_shape: 512
sub_name: 'arc_res50'
backbone_type: 'ResNet50' # or 'MobileNetV2'
head_type: ArcHead # or 'NormHead': FC to targets.
is_ccrop: False # central-cropping or not
# train
train_dataset: './data/ms1m_bin.tfrecord' # or './data/ms1m.tfrecord'
binary_img: True # False if dataset is online decoding
num_classes: 85742
num_samples: 5822653
epochs: 5
base_lr: 0.01
w_decay: !!float 5e-4
save_steps: 1000
# test
test_dataset: '/your/path/to/test_dataset'
```
Note:
- The `sub_name` is the name of outputs directory used in checkpoints and logs folder. (make sure of setting it unique to other models)
- The `head_type` is used to choose [ArcFace](https://arxiv.org/abs/1801.07698) head or normal fully connected layer head for classification in training. (see more detail in [./modules/models.py](https://github.com/peteryuX/arcface-tf2/blob/master/modules/models.py#L90-L94))
- The `is_ccrop` means doing central-cropping on both trainging and testing data or not.
- The `binary_img` is used to choose the type of training data, which should be according to the data type you created in the [Data-Preparing](#Data-Preparing).
### Training
Here have two modes for training your model, which should be perform the same results at the end.
```bash
# traning with tf.GradientTape(), great for debugging.
python train.py --mode="eager_tf" --cfg_path="./configs/arc_res50.yaml"
# training with model.fit().
python train.py --mode="fit" --cfg_path="./configs/arc_res50.yaml"
```
### Testing
You can download my trained models for testing from [Benchmark and Models](#Benchmark-and-Models) without training it yourself. And, evaluate the models you got with the corresponding cfg file on the testing dataset. The testing code in [./modules/evaluations.py](https://github.com/peteryuX/arcface-tf2/blob/master/modules/evaluations.py) were modified from [face.evoLVe](https://github.com/ZhaoJ9014/face.evoLVe.PyTorch).
```bash
python test.py --cfg_path="./configs/arc_res50.yaml"
```
### Encode Input Image
You can also encode image into latent vector by model. For example, encode the image from [./data/BruceLee.jpg](https://github.com/peteryuX/arcface-tf2/blob/master/data/BruceLee.jpg) and save to `./output_embeds.npy` as following.
```bash
python test.py --cfg_path="./configs/arc_res50.yaml" --img_path="./data/BruceLee.jpg"
```
****
## Benchmark and Models
:coffee:
Verification results (%) of different backbone, head tpye, data augmentation and loss function.
| Backbone | Head | Loss | CCrop | LFW | AgeDB-30 | CFP-FP | Download Link |
|----------|------|------|-------|-----|--------|----------|---------------|
| [ResNet50](https://arxiv.org/abs/1512.03385) | [ArcFace](https://arxiv.org/abs/1801.07698) | Softmax | False | 99.42 | 95.32 | 92.56 | [GoogleDrive](https://drive.google.com/file/d/1HasWQb86s4xSYy36YbmhRELg9LBmvhvt/view?usp=sharing) |
| [MobileNetV2](https://arxiv.org/abs/1801.04381) | [ArcFace](https://arxiv.org/abs/1801.07698) | Softmax | False | 99.13 | 91.62 | 91.50 | [GoogleDrive](https://drive.google.com/file/d/1qG8BChcPHzKuGwjJhrpeIxBqQmhpLvTX/view?usp=sharing) |
| [ResNet50](https://arxiv.org/abs/1512.03385) | [ArcFace](https://arxiv.org/abs/1801.07698) | Softmax | True | 99.38 | 95.13 | 94.87 | [GoogleDrive](https://drive.google.com/file/d/1zUulC-4hSY_kPqZpcoIHO96OmjMivuKB/view?usp=sharing) |
| [MobileNetV2](https://arxiv.org/abs/1801.04381) | [ArcFace](https://arxiv.org/abs/1801.07698) | Softmax | True | 98.88 | 91.58 | 93.19 | [GoogleDrive](https://drive.google.com/file/d/1nSnIc0eV0MkSjg48x29PJwTt3fGXKDU4/view?usp=sharing) |
Note:
- The 'CCrop' tag above means doing central-cropping on both trainging and testing data, which could eliminate the redundant boundary of intput face data (especially for AgeDB-30).
- All training settings of the models can be found in the corresponding [./configs/*.yaml](https://github.com/peteryuX/arcface-tf2/tree/master/configs) files.
- **Based on the property of the training dataset, all the pre-trained models can only be used for non-commercial applications.**
****
## References
:hamburger:
Thanks for these source codes porviding me with knowledges to complete this repository.
- https://github.com/deepinsight/insightface/tree/master/recognition (Official)
- Face Analysis Project on MXNet http://insightface.ai
- https://github.com/zzh8829/yolov3-tf2
- YoloV3 Implemented in TensorFlow 2.0
- https://github.com/ZhaoJ9014/face.evoLVe.PyTorch
- face.evoLVe: High-Performance Face Recognition Library based on PyTorch
- https://github.com/luckycallor/InsightFace-tensorflow
- Tensoflow implementation of InsightFace (ArcFace: Additive Angular Margin Loss for Deep Face Recognition).
- https://github.com/dmonterom/face_recognition_TF2
- Training a face Recognizer using ResNet50 + ArcFace in TensorFlow 2.0