BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260114T163633Z
LOCATION:Darling Harbour Theatre\, Level 2 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T093000
DTEND;TZID=Australia/Melbourne:20231212T124500
UID:siggraphasia_SIGGRAPH Asia 2023_sess209_papers_346@linklings.com
SUMMARY:ProSpect: Prompt Spectrum for Attribute-Aware Personalization of D
 iffusion Models
DESCRIPTION:Yuxin Zhang (MAIS, Institute of Automation, Chinese Academy of
  Sciences; School of Artificial Intelligence, University of Chinese Academ
 y of Sciences); Weiming Dong (MAIS, Institute of Automation, Chinese Acade
 my of Sciences; School of AI,University of Chinese Academy of Sciences); F
 an Tang (Institute of Computing Technology, Chinese Academy of Sciences); 
 Nisha Huang (School of AI,University of Chinese Academy of Sciences; MAIS,
  Institute of Automation, Chinese Academy of Sciences); Haibin Huang and C
 hongyang Ma (Kuaishou Technology); Tong-Yee Lee (National Cheng-Kung Unive
 rsity); Oliver Deussen (University of Konstanz); and Changsheng Xu (MAIS, 
 Institute of Automation, Chinese Academy of Sciences; School of Artificial
  Intelligence, University of Chinese Academy of Sciences)\n\nPersonalizing
  generative models offers a way to guide image generation with user-provid
 ed references. Current personalization methods can invert an object or con
 cept into the textual conditioning space and compose new natural sentences
  for text-to-image diffusion models. However, representing and editing spe
 cific visual attributes like material, style, layout, etc. remains a chall
 enge, leading to a lack of disentanglement and editability. To address thi
 s problem, we propose a novel approach that leverages the step-by-step gen
 eration process of diffusion models, which generate images from low- to hi
 gh-frequency information, providing a new perspective on representing, gen
 erating, and editing images.  We develop Prompt Spectrum Space, an expande
 d textual conditioning space, and a new image representation method called
  ProSpect. ProSpect represents an image as a collection of inverted textua
 l token embeddings encoded from per-stage prompts, where each prompt corre
 sponds to a specific generation stage (i.e., a group of consecutive steps)
  of the diffusion model. Experimental results demonstrate that ProSpect of
 fers better disentanglement and controllability compared to existing metho
 ds. We apply ProSpect in various personalized attribute-aware image genera
 tion applications, such as image-guided or text-driven manipulations of ma
 terials, style, and layout, achieving previously unattainable results from
  a single image input without fine-tuning the diffusion models. Code: \url
 {github.com/zyxElsa/ProSpect}.\n\nRegistration Category: Full Access, Enha
 nced Access, Trade Exhibitor, Experience Hall Exhibitor\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_346&sess=sess209
END:VEVENT
END:VCALENDAR