Our real life datasets are all shot in-house and start with each model signing a GDPR compliant biometric release. Our highly engineered process was designed by data scientists, and all our datasets have been injected with appropriate labeling noting emotional state, age, gender, ethnicity and incorporate technical aspects such as angles or perspectives embedded into universally accepted IPTC metadata. Each image is a 42MP masterpiece of information rich pixels and can be cropped or reduced as wished.
Our datasets offer users full legal security and perform up to 10 times better than dirty datasets scraped from the net.
Diversity is baked into our product – as unbiased as it gets.