WFLW Dataset

Estimated reading: 4 minutes

Visualization of the WFLW dataset in the Deep Lake UI

WFLW dataset

What is WFLW Dataset?

The WFLW dataset comprises 98 entirely hand-annotated landmarks. This dataset contains extensive attribute annotations such as occlusion, position, make-up, lighting, blur, and expression to allow for a more thorough examination of each face.

Faces in the proposed dataset bring considerable changes in expression, posture, and occlusion when compared to the prior dataset. Instead of switching between several evaluation techniques in different datasets, we may simply test the robustness of posture, occlusion, and expression on the suggested dataset.

Download WFLW Dataset in Python

Instead of downloading the WFLW dataset in Python, you can effortlessly load it in Python via our Deep Lake open-source with just one line of code.

Load WFLW Dataset Training Subset in Python

				
					import deeplake
ds = deeplake.load("hub://activeloop/wflw-train")

Load WFLW Dataset Testing Subset in Python

				
					import deeplake
ds = deeplake.load("hub://activeloop/wflw-test")

WFLW Dataset Structure

WFLW Data Fields

image: tensor containing the image.
boxes: tensor containing bounding box array
keypoints: tensor to represent 98 key points on the face.
makeups: tensor to detect the presence of makeup on the face
occlusions: tensor to detect any occlusion for the facial image
blurs: tensor to identify if the image is blurred
expressions: tensor to identify if there is an expression on the facial image
illuminations: tensor to identify if there is illumination in the image
poses: tensor to identify if there is any pose present in the face

WFLW Data Splits

WFLW training split comprises 7500 images.
WFLW testing split comprises 2500 images.

How to use WFLW Dataset with PyTorch and TensorFlow in Python

Train a model on WFLW dataset with PyTorch in Python

Let’s use Deep Lake built-in PyTorch one-line dataloader to connect the data to the compute:

				
					dataloader = ds.pytorch(num_workers=0, batch_size=4, shuffle=False)

Train a model on WFLW dataset with TensorFlow in Python

				
					dataloader = ds.tensorflow()

Additional Information about WFLW Dataset

WFLW Dataset Description

Homepage: https://wywu.github.io/dataset/
Paper: Deep Entwined Learning Head Pose and Face Alignment Inside an Attentional Cascade with Doubly-Conditional fusion Arnaud Dapogny, Kévin Bailly, Matthieu Cord
Point of Contact: [email protected]

WFLW Dataset Curators

Arnaud Dapogny, Kévin Bailly, Matthieu Cord

WFLW Dataset Licensing Information

Deep Lake users may have access to a variety of publicly available datasets. We do not host or distribute these datasets, vouch for their quality or fairness, or claim that you have a license to use the datasets. It is your responsibility to determine whether you have permission to use the datasets under their license.

If you’re a dataset owner and do not want your dataset to be included in this library, please get in touch through a GitHub issue. Thank you for your contribution to the ML community!

WFLW Dataset Citation Information

				
					@inproceedings{wayne2018lab,
 author = {Wu, Wayne and Qian, Chen and Yang, Shuo and Wang, Quan and Cai, Yici and Zhou, Qiang},
 title = {Look at Boundary: A Boundary-Aware Face Alignment Algorithm},
 booktitle = {CVPR},
 month = June,
 year = {2018}
}

WFLW Dataset FAQs

What is the WFLW dataset for Python?

The WFLW (Wider Facial Landmarks in the Wild) database comprises 10000 images of faces (7500 for training and 2500 for testing) with 98 annotated landmarks. The database features attribute annotations such as occlusion, head pose, make-up, illumination, blur, and expressions. The WFLW dataset introduces large variations in expression, pose, and occlusion.

What is the WFLW dataset used for?

The WFLW dataset was created to enable future research on boundary-aware face alignment. This dataset was introduced by Wu et al. in the Look at Boundary: A Boundary-Aware Face Alignment Algorithm paper. The dataset allows you to unify training and testing across different factors, such as poses, expressions, illuminations, makeups, occlusions, and blurriness.

How to download the WFLW dataset in Python?

You can load the WFLW dataset with one line of code using the open-source package Activeloop Deep Lake in Python. See detailed instructions on how to load the WFLW dataset training subset or WFLW dataset testing subset in Python.

How can I use the WFLW dataset in PyTorch or TensorFlow?

You can stream WFLW while training a model in PyTorch or TensorFlow with one line of code using the open-source package Activeloop Deep Lake in Python. See detailed instructions on how to train a model on the WFLW dataset with PyTorch in Python or train a model on the WFLW dataset with TensorFlow in Python.