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:3008\, 3rd Floor
DTSTART;TZID=America/Los_Angeles:20240626T114500
DTEND;TZID=America/Los_Angeles:20240626T120000
UID:dac_DAC 2024_sess126_RESEARCH1267@linklings.com
SUMMARY:Auto-ISP: An Efficient Real-Time Automatic Hyperparameter Optimiza
 tion Framework for ISP Hardware System
DESCRIPTION:Research Manuscript\n\nJiaming Liu (Fudan University); Zihao L
 iu (Alibaba Group); Xuan Huang, Ruoxi Zhu, Qi Zheng, and Zhijian Hao (Fuda
 n University); Tao Liu (Lawrence Technological University); and Jun Tao an
 d Yibo Fan (Fudan University)\n\nImage Signal Processor (ISP) is widely us
 ed in intelligent edge devices across various scenarios. The intricate and
  time-consuming tuning process demands substantial expertise. Current AI-b
 ased auto-tuning operates discretely offline, relying on predefined scenes
  with human intervention, leading to inconvenient manipulation, with poten
 tially fatal impacts on downstream tasks in unforeseen scenes. We propose 
 a real-time automatic hyperparameter optimization ISP hardware system to a
 ddress real-world scenarios. Our design features a tri-step framework and 
 a hardware accelerator, demonstrating superior performance in human and co
 mputer vision tasks, even in real-time unforeseen scenes. Experiments show
 case its practicality, achieving 1080P@75FPS/240FPS in FPGA/ASIC, respecti
 vely.\n\nTopic: AI, Design\n\nKeyword: AI/ML System and Platform Design\n\
 nSession Chair: Hsien-Hsin Sean Lee (Intel Corporation)
END:VEVENT
END:VCALENDAR