Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/joonb14/jhface

Face Recognition training and testing framework with tensorflow 2.0 based on the well implemented arcface-tf2. Changes are added to provide tensorflow lite conversion, and provide additional backbones, loss functions.
https://github.com/joonb14/jhface

arcface arcface-loss arcface-tf2 arcface-tflite cosface cosface-loss cosface-tflite efficientnet-lite face-recognition face-recognition-python face-recognition-tensorflow ipynb mobilenetv3 python tensorflow-lite tensorflow2 tensorflowlite tf2 tflite

Last synced: 28 days ago
JSON representation

Face Recognition training and testing framework with tensorflow 2.0 based on the well implemented arcface-tf2. Changes are added to provide tensorflow lite conversion, and provide additional backbones, loss functions.

Awesome Lists containing this project

README 

        
# JHFace



This project is based on the well implemented [arcface-tf2](https://github.com/peteryuX/arcface-tf2).

The things that rise error when converting the model to tflite was changed.



### Training the Model

For preparing data, follow the instructions provided in  [data-preparing](https://github.com/peteryuX/arcface-tf2#data-preparing)

Then checkout the [TensorFlow ArcFace.ipynb](https://github.com/joonb14/JHFace/blob/main/TensorFlow%20ArcFace.ipynb).



##### Backbones w/ ImageNet pretrained weights:



NasNet case, because of this [issue](https://github.com/keras-team/keras-applications/issues/78), we manually download the weight file and explicitly load it in [models.py](https://github.com/joonb14/JHFace/blob/main/models.py) file. We tried to provide pretrained weights for the MobileNet and EfficientNet models. However the official Keras implementations of the [MobileNetV3 has built in preprocessing layers inside the model](https://github.com/tensorflow/tensorflow/pull/47808#pullrequestreview-612848161). Also this accounts to [EfficientNet as well](https://github.com/tensorflow/tensorflow/pull/48276). So for the EfficientNet and EfficientNetLite, we used [Weights Transfer.ipynb](https://github.com/joonb14/JHFace/blob/main/Weights%20Transfer.ipynb) for extracting pretrained weights. To use the pretrained weights for NasNet, EfficientNet, EfficientNetLite **please download the weights using this [link](https://drive.google.com/file/d/1EriCfISIfeRAOso1DRRuQD8TFd_as35n/view?usp=sharing).** Then unzip it inside the path/to/JHFace/weights/ directory. All of the pretrained weights are provided by(or extracted from) [tf.keras.applications](https://www.tensorflow.org/api_docs/python/tf/keras/applications)



* MobileNet

* MobileNetV2

* InceptionResNetV2

* InceptionV3

* ResNet50

* ResNet50V2

* ResNet101V2

* NASNetLarge

* NASNetMobile

* Xception

* MobileNetV3Large

* MobileNetV3Small

* EfficientNetLite0 ~ Lite6

* EfficientNetB0 ~ B7



##### Backbones w/o ImageNet pretrained weights:



We implemented MnasNet models looking at the [official code](https://github.com/tensorflow/tpu/blob/master/models/official/mnasnet/mnasnet_model.py). If there's a bug, please tell us through the github issue page!



* MnasNetA1

* MnasNetB1

* MnasNetSmall



##### Loss Function



* [ArcFace](https://openaccess.thecvf.com/content_CVPR_2019/html/Deng_ArcFace_Additive_Angular_Margin_Loss_for_Deep_Face_Recognition_CVPR_2019_paper.html)

* [CosFace](https://openaccess.thecvf.com/content_cvpr_2018/html/Wang_CosFace_Large_Margin_CVPR_2018_paper.html)

* [SphereFace](https://openaccess.thecvf.com/content_cvpr_2017/papers/Liu_SphereFace_Deep_Hypersphere_CVPR_2017_paper.pdf)

##### Configuration

in the [TensorFlow ArcFace.ipynb](https://github.com/joonb14/JHFace/blob/main/TensorFlow%20ArcFace.ipynb), we provided simple configuration values. To change the model backbone, just change the backbone_type parameter. To change the loss function, just change the head_type parameter.

```python

### MS1M dataset



batch_size = 128 # Initially 128

input_size = 112

embd_shape = 512

head_type = 'ArcHead' # 'ArcHead', 'CosHead', 'SphereHead'

# Backbones w/ pretrained weights:

#     MobileNet, MobileNetV2, InceptionResNetV2, InceptionV3, ResNet50, ResNet50V2, ResNet101V2, NASNetLarge, NASNetMobile, Xception, MobileNetV3Large, MobileNetV3Small, EfficientNetLite0~6, EfficientNetB0~7

# Backbones w/o pretrained weights:

#      MnasNetA1, MnasNetB1, MnasNetSmall 

backbone_type = 'EfficientNetLite0' 

w_decay=5e-4

num_classes = 85742 

dataset_len = 5822653 

if head_type == 'SphereHead':

    base_lr = 0.01 

    margin = 1.35

    logist_scale = 30.0 

elif head_type == 'CosHead':

    base_lr = 0.01 

    margin=0.35

    logist_scale=64

elif head_type == 'ArcHead':

    base_lr = 0.01 

    margin=0.5

    logist_scale=64

else:

    base_lr = 0.01 # initially 0.01

epochs = 20

save_steps = 1000

train_size = int(0.8 * dataset_len)

print(train_size)

steps_per_epoch = train_size // batch_size

print(steps_per_epoch)

val_size = dataset_len - train_size

print(val_size)

validation_steps = val_size // batch_size

print(validation_steps)

steps = 1

is_ccrop=False

binary_img=True

is_Adam = False

```



## Converting the Model to TensorFlow Lite



checkout the [TFLite conversion.ipynb](https://github.com/joonb14/arcface-tflite/blob/main/TFLite%20conversion.ipynb).

Int8 quantization is supported, we checked with MobileNetV2 and EfficientNet-lite0.



## For Face Verification

We downloaded the testing dataset from [here.](https://github.com/peteryuX/arcface-tf2#testing-dataset)

With the data use the [verification.ipynb](https://github.com/joonb14/JHFace/blob/main/verification.ipynb) for verification test.



## For Face Identificaiton



checkout the [Face Identification with Centroid Vector.ipynb](https://github.com/joonb14/JHFace/blob/main/Face%20Identification%20with%20Centroid%20Vector.ipynb).