BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Unifying GstAnalytics Metadata: Achieving Cross-Vendor Interoperab
 ility in Video Analytics
DTSTART;VALUE=DATE-TIME:20261010T082000Z
DTEND;VALUE=DATE-TIME:20261010T090000Z
DTSTAMP;VALUE=DATE-TIME:20261009T233018Z
UID:indico-contribution-587@indico.freedesktop.org
DESCRIPTION:Speakers: Oskar Fiedot (Intel)\, Pep Cots (Milestone Systems)\
 , Daniel Morin (Collabora)\nAs the GStreamer framework continues to be a l
 eading tool for building complex video analytic pipelines\, the need for s
 eamless interoperability across vendor components is more critical than ev
 er. While the GstAnalytics library\, introduced in GStreamer 1.24\, provid
 es a flexible foundation for representing AI-driven metadata\, its very fl
 exibility has resulted in slight differences between various implementatio
 ns\, hindering the creation of truly plug-and-play analytics pipelines.\n\
 nThis talk presents a collaborative effort by Intel\, Collabora\, and Mile
 stone Systems to develop a unified convention for GstAnalytics metadata\, 
 ensuring consistent interpretation and consumption across the ecosystem. W
 e will introduce a comprehensive set of guidelines that standardize the us
 e of GstAnalytics relations for linking object detections\, classification
 s\, keypoints\, and segmentation data.\nDaniel Morin will present key chal
 lenge in this effort: striking the right balance between specificity and g
 enerality\, where the convention needed to be concrete enough to guarantee
  interoperability between independently developed component\, yet general 
 enough to avoid over-fitting to a specific industry use-cases.\n\nTo demon
 strate the power of this standardized approach\, we will cover two signifi
 cant industry adoptions:\n1. Intel's DL Streamer: We will showcase how Int
 el has transitioned its DL Streamer framework from custom metadata formats
  to this new GstAnalytics convention\, enabling broader compatibility with
  the GStreamer ecosystem.\n2. Milestone Systems' License Plate Recognition
  Integration Case Study: We will present a practical development use case 
 exploring the challenges and lessons learned while integrating analytics c
 omponents from multiple vendors. By combining components from Intel DL Str
 eamer and Milestone within the same container\, we will discuss the requir
 ements and design decisions involved in enabling analytics metadata produc
 ed by one vendor to be consumed by another.\n\nThrough a demonstration (re
 corded video) of a complete License Plate Recognition (LPR) solution using
  the pipeline vmssrc → gvamotiondetect → vmslpr → gvatrack → vmsme
 tadataexpose\, we will show how GstAnalytics metadata serves as a common l
 anguage between elements.\nAttendees will leave with a clear understanding
  of the new GstAnalytics proposed conventions\, the integration challenges
  encountered\, and see firsthand how industry leaders are building the nex
 t generation of seamless\, cross-vendor video analytics workflows.\n\nhttp
 s://indico.freedesktop.org/event/14/contributions/587/
LOCATION:
URL:https://indico.freedesktop.org/event/14/contributions/587/
END:VEVENT
END:VCALENDAR