The number of human-related applications of computer vision and AI is growing by the day. From medicine to retail to manufacturing and security, AI-powered solutions are becoming more prevalent but one constant remains – the need for high-quality data for training AI models. Face analysis-related use cases intensify and complicate this need even more, for example, face recognition, gaze estimation, face segmentation, facial expressions analysis, etc. Sky Engine AI is challenging this reality with their approach to synthetic data for training computer vision AI models.

Why train computer vision models on synthetic data?

The current and most popular method of using manually labelled real-world images to train deep learning models in computer vision has multiple drawbacks. They include:

• Legal (and ethical) concerns, as privacy regulations, such as GDPR and similar, prevent the use of most of existing data with real human faces (labelled and unlabelled). Synthetic data is free of these constraints.
• Labelling issues caused by manual or semi-manual labelling which is often imprecise and incomplete.
• Low diversity of datasets. Face recognition AI models do not work properly in real life when trained and tested on datasets lacking in various accessories, hair styles, age, gender, skin colour, etc., or on unwell-balanced data collections.
• Context bias. Training outcomes obtained with real-world datasets don’t accurately reflect reality – a seemingly illogical effect, but easy to explain. Some cases are so rare that obtaining a large number of images representing that case would last 100 years. Other cases are related to contamination with outside artifacts due to an insufficient number of labelled images.
• Lack of 3D information in ground truth. Labelling farms are not able to add information such as 3D annotations, 3D key points, 3D bounding boxes, and depth maps to the real-world images.

Sky Engine AI’s new synthetic dataset ‘The Face It!–6M’ looks to solve these problems by creating a dataset collection of synthetic human characters, created especially for Vision AI tasks, that enables clear outcomes of training computer vision AI models.