Dark Mode
Facial Landmark Training Data
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset is designed for facial keypoint detection, providing the necessary data to identify the location of landmarks on a human face. It is particularly useful for those keen on developing their own models for facial landmark identification. Detecting these keypoints is a complex task due to significant variations in 3D pose, size, position, viewing angle, and illumination conditions. The dataset facilitates the advancement of computer vision research in addressing these challenges. Each keypoint is precisely defined by an (x,y) real-valued pair, representing pixel indices.
Columns
The dataset contains 30 columns, representing the X and Y pixel coordinates for 15 distinct facial keypoints. These keypoints include:
- left_eye_center_x, left_eye_center_y: Coordinates for the centre of the left eye.
- right_eye_center_x, right_eye_center_y: Coordinates for the centre of the right eye.
- left_eye_inner_corner_x, left_eye_inner_corner_y: Coordinates for the inner corner of the left eye.
- left_eye_outer_corner_x, left_eye_outer_corner_y: Coordinates for the outer corner of the left eye.
- right_eye_inner_corner_x, right_eye_inner_corner_y: Coordinates for the inner corner of the right eye.
- right_eye_outer_corner_x, right_eye_outer_corner_y: Coordinates for the outer corner of the right eye.
- left_eyebrow_inner_end_x, left_eyebrow_inner_end_y: Coordinates for the inner end of the left eyebrow.
- left_eyebrow_outer_end_x, left_eyebrow_outer_end_y: Coordinates for the outer end of the left eyebrow.
- right_eyebrow_inner_end_x, right_eyebrow_inner_end_y: Coordinates for the inner end of the right eyebrow.
- right_eyebrow_outer_end_x, right_eyebrow_outer_end_y: Coordinates for the outer end of the right eyebrow.
- nose_tip_x, nose_tip_y: Coordinates for the tip of the nose.
- mouth_left_corner_x, mouth_left_corner_y: Coordinates for the left corner of the mouth.
- mouth_right_corner_x, mouth_right_corner_y: Coordinates for the right corner of the mouth.
- mouth_center_top_lip_x, mouth_center_top_lip_y: Coordinates for the centre of the top lip.
- mouth_center_bottom_lip_x, mouth_center_bottom_lip_y: Coordinates for the centre of the bottom lip. "Left" and "right" refer to the subject's point of view.
Distribution
The dataset is provided in a CSV file,
training.csv, which is approximately 1.52 MB in size. It contains 30 columns. The CSV records are linked to corresponding images via an index, where index.jpg in a training folder represents the image for a record. The dataset includes around 7000 records for nose tip and eye centre keypoints, while other keypoints like inner/outer eye corners, eyebrows, and mouth corners have approximately 2200-2300 valid records, with a notable percentage of missing values for these specific keypoints.Usage
This dataset is ideal for training and evaluating machine learning models focused on:
- Facial Keypoint Detection: Building models to automatically locate specific landmarks on faces.
- Facial Recognition Systems: Enhancing the precision of facial identification and verification.
- Emotion Detection: Analysing facial expressions through keypoint movements.
- Augmented Reality (AR) Applications: Developing filters and effects that accurately overlay on faces.
- Human-Computer Interaction: Enabling gesture recognition and gaze tracking.
Coverage
The dataset focuses on the coordinates of facial keypoints. Specific geographic regions, time ranges, or detailed demographic information about the subjects is not available within the provided material. The data generally pertains to human faces, with the challenges highlighted due to variations in 3D pose, size, position, viewing angle, and illumination.
License
CC0: Public Domain
Who Can Use It
This dataset is suitable for:
- Computer Vision Researchers: To advance algorithms for facial landmarking and analysis.
- Machine Learning Engineers: For training deep learning models like Convolutional Neural Networks (CNNs) for image-based tasks.
- Data Scientists: To explore techniques for feature extraction and pattern recognition on facial data.
- Developers: Creating applications requiring accurate face tracking or augmentation.
Dataset Name Suggestions
- Facial Keypoint Detection Dataset
- Human Face Landmark Coordinates
- Image-based Facial Feature Dataset
- Face Keypoint Localisation Data
- Facial Landmark Training Data
Attributes
Original Data Source: Facial Landmark Training Data
QUALITY