Alongside a number of new announcements surrounding Microsoft's AI initiative, Project Oxford, Microsoft introduced a number of updates to its Project Oxford Face APITrack this API. New features included in the Face API updates include facial hair and smile prediction tools, as well as improved visual age estimation and gender identification. The facial hair tool allows users to rate facial hair and the smile prediction tool enhances the Project's existing portfolio of recognizing emotions and facial features.
Additional announcements included public beta releases of spell check, video, speaker recognition, and custom recognition intelligence services (CRIS) tools. The spell check tool can be integrated into mobile and cloud-based apps and includes common slang words (e.g. "gonna") and difficult to spot errors (e.g. "four" vs. "for"). The video tool enhances editing ability through features like face tracking, motion detection, and stabilization. Speaker recognition identifies the person speaking because individuals all maintain unique voices. The CRIS tool customizes speech recognition for difficult environments (e.g. noisy public spaces, shop floors, etc.).
Project Oxford aims to arm developers with machine learning tools that replicate recognition and response actions previously limited to humans. While traditional computers struggled with basic human skills (e.g. detecting a smile), Project Oxford continues to improve machine learning. API access to such tools enables developers to include machine learning capabilities into third party mobile and cloud-based apps. For a complete update on new Project Oxford updates, visit the project site.