FASSEG: A FAce semantic SEGmentation repository for face image analysis.

Area of Science:

• Computer Vision
• Image Processing
• Machine Learning

Background:

• Automated face segmentation is crucial for various applications.
• Existing datasets may lack diversity in poses and annotation precision.

Purpose of the Study:

• Introduce the FASSEG repository for face segmentation research.
• Provide diverse datasets for training and evaluating face segmentation models.

Main Methods:

• The FASSEG repository comprises four subsets: frontal01, frontal02, frontal03, and multipose01.
• Datasets include RGB images with varying ground-truth mask precision.
• Annotations cover six classes: mouth, nose, eyes, hair, skin, and background.

Main Results:

• Frontal subsets offer varying levels of annotation detail.
• Multipose subset provides extensive multi-pose face data.
• All subsets are suitable for training and testing face segmentation algorithms.

Conclusions:

• FASSEG repository addresses the need for comprehensive face image data.
• Facilitates advancements in automated face segmentation techniques.
• Supports research in diverse facial imaging conditions.