GitHub - artslidd/FaceBoxes.keras: Keras implementation of FaceBoxes

FaceBoxes implementation in Keras

Status: just started, fixing bugs, wanting helps

This is a Keras port of the FaceBoxes architecture introduced by Shifeng Zhang et al. in the paper FaceBoxes: A CPU Real-time Face Detector with High Accuracy.

The project refers to the SSD framework: SSD_keras

The first implementation is without density.

The faceboxes code are written in ./models/FaceBoxes.py. The training process and the setting of parameters are in ./FaceBoxes_train.py.

Name		Name	Last commit message	Last commit date
Latest commit History 10 Commits
bounding_box_utils		bounding_box_utils
data_generator		data_generator
eval_utils		eval_utils
examples		examples
keras_layers		keras_layers
keras_loss_function		keras_loss_function
misc_utils		misc_utils
models		models
saved_model		saved_model
ssd_encoder_decoder		ssd_encoder_decoder
training_summaries		training_summaries
.gitattributes		.gitattributes
CONTRIBUTING.md		CONTRIBUTING.md
ISSUE_TEMPLATE.md		ISSUE_TEMPLATE.md
LICENSE.txt		LICENSE.txt
README.md		README.md
__init__.py		__init__.py
faceboxes_framework.jpg		faceboxes_framework.jpg
faceboxes_train.py		faceboxes_train.py
log.txt		log.txt
result1.jpg		result1.jpg
ssd300_evaluation.ipynb		ssd300_evaluation.ipynb
ssd300_evaluation_COCO.ipynb		ssd300_evaluation_COCO.ipynb
ssd300_inference.ipynb		ssd300_inference.ipynb
ssd300_inference.py		ssd300_inference.py
ssd300_pascal_07+12_training_log.csv		ssd300_pascal_07+12_training_log.csv
ssd300_training.ipynb		ssd300_training.ipynb
ssd512_inference.ipynb		ssd512_inference.ipynb
ssd7_training.ipynb		ssd7_training.ipynb
weight_sampling_tutorial.ipynb		weight_sampling_tutorial.ipynb