NEXYTE Product Description Page 18 Text-inferred Relations 4.2.2 Image and Video Analytics NEXYTE applies ML-powered image and video analytics to enrich media content with intelligence extracted from the analyzed media. The following image analysis and intelligence extraction abilities are supported for image and video media files, per specific capability. + Face detection & recognition detects multiple faces within a picture or video image, along with key related facial attributes such as gender, age, and ethnicity. Using this data, the system links image/video files to existing entities, provides possible matches, and allows users to identify entities in stored media files, or find persons that match specific facial characteristics. Applies to both image and video files. + Object tagging detects objects such as weapons (knife, gun, etc.), vehicles, bags, or other objects in accordance with the requirements of the customer. Applies to both image and video files. + Optical Character Recognition detects and extracts text from an image, supporting multiple languages. Applies to image files only. + Transcription of video file audio content extracts and indexes audio content from the video file allowing translation and text-based searches for video content. + Image-inferred relations detects and categorizes relations between persons based on their presence in the image and level of intimacy (hugging, holding hands). Applies to image files only. 4.2.3 Audio Analytics NEXYTE applies ML powered audio analytics to ingested audio content to enrich audio content with intelligence metadata and related insights: + Audio language identification – Automatic identification of languages spoken in the audio content. + Gender identification – Automatic tagging of speakers’ gender within the audio recording