BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240626T180034Z
LOCATION:Level 2 Lobby
DTSTART;TZID=America/Los_Angeles:20240626T180000
DTEND;TZID=America/Los_Angeles:20240626T190000
UID:dac_DAC 2024_sess256_LBR101@linklings.com
SUMMARY:Late Breaking Results: A real-time diffusion-based filter for huma
 n pose estimation on edge devices
DESCRIPTION:Late Breaking Results Poster\n\nChiara Bozzini, Michele Boldo,
  Enrico Martini, and Nicola Bombieri (University of Verona)\n\nHuman Pose 
 Estimation (HPE) is increasingly being adopted in a wide range of applicat
 ions, from healthcare to Industry 5.0. To address the intrinsic inaccuracy
  of such CNN-based software, the current trend involves applying filtering
  models to refine and improve the inference results. However, state-of-the
 -art filtering models are computationally intensive, limiting their use in
  resource-constrained devices. To overcome this limitation, we propose a r
 eal-time filtering technique based on diffusion models designed specifical
 ly for edge devices. Through a micro-benchmarking phase, we analyze how th
 e model responds to various levels of noise and select the optimal setup f
 or specific application scenarios. Using a widely available edge device, w
 e evaluated the model's performance on both synthetic and real noise gener
 ated by a state-of-the-art HPE system. Preliminary results demonstrate a s
 ignificant improvement in real-time filtering performance with minimal com
 putational overhead.
END:VEVENT
END:VCALENDAR