Уровень 0 · материалов: 3
В кластер входят документы о технологиях анализа, идентификации и распознавания лиц с помощью ИИ, но не инструменты для создания цифровых аватаров в реальном времени.
Общие признаки: распознавание лиц, искусственный интеллект, анализ мимики и внешности, компьютерное зрение
Группа выше: Распознавание лиц и объектов
Смысл: The text describes Facebook's creation of the DeepFace algorithm, which uses 3D modeling to recognize faces with near-human accuracy, potentially automating social media tagging and advancing AI.
Facebook developed DeepFace, a facial recognition algorithm with 97.25% accuracy that uses 3D modeling to identify people from various angles.
Смысл: The text proposes a hybrid platform (AI-driven and crowdsourced) to find people with similar physical appearances, arguing that human perception is necessary to supplement the inaccuracies of current facial recognition technology.
The author suggests creating a website that finds visual look-alikes using a combination of facial recognition algorithms and user-submitted links.
Смысл: The main idea is that computer vision and AI (specifically Microsoft Face API) can transform offline business intelligence by replacing unreliable manual surveys with objective, real-time data on audience demographics and emotional reactions.
CVizi uses Microsoft Azure's Face API and specialized hardware to provide businesses with objective analytics on the age, gender, and emotions of offline audiences in cinemas, events, and stores.