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:20260730T152727Z
LOCATION:Mtg Room 101A
DTSTART;TZID=America/Los_Angeles:20260729T142100
DTEND;TZID=America/Los_Angeles:20260729T143400
UID:dac_DAC 2026_sess126_RESEARCH3067@linklings.com
SUMMARY:Scale-Gest: Scalable Model-Space Synthesis and Runtime Selection f
 or On-Device Gesture Detection
DESCRIPTION:Abdul Basit, Saim Rehman, and Muhammad Shafique (New York Univ
 ersity, Abu Dhabi)\n\nRealizing on-device ML-based gesture detection under
  tight real-time performance, energy and memory constraints is challenging
 , especially when considering mobile devices with varying battery-power le
 vels. Existing EdgeAI deployments typically rely on a single fixed detecto
 r, limiting optimization opportunities. We present Scale-Gest, a novel run
 time-adaptive gesture detection framework that expands the detector space 
 into a dense family of tiny-YOLO architectures. We introduce multiple nove
 l device-calibrated ACE (Accuracy-Complexity-Energy) profiles by analyzing
  different model-resolution-stride operating points. A run-time controller
  selects an appropriate ACE mode under user-defined and battery-constraint
 s, while a motion-aware hand-gesture-tracking ROI gate crops the input for
  reduced-complexity detection. To evaluate performance of our system in re
 al-world car driving scenarios, we introduce a temporally-annotated Driver
  Simulated Gesture (DSG-18) dataset. Scale-Gest maintains event-level F1 w
 hile significantly reducing energy and latency compared to single-detector
  approaches.  On a battery-powered laptop running gesture streams, our ACE
  controller reduces per-frame energy by ≈4× (from ≈6.9 mJ to ≈1.6 mJ) whil
 e maintaining high gesture-detection performance (event-level F1 ≈ 0.8–0.9
 ) and low mean latency (≈6 ms).\n\nTopics: AI\n\nTrack: AI3-II. AI/ML Appl
 ication and Infrastructure\n\nSession Chairs: Yingyan (Celine) Yin (Georgi
 a Institute of Technology) and Yan Fang (Kennesaw State University)\n\n
END:VEVENT
END:VCALENDAR