BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260817T171534Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241203T132300
DTEND;TZID=Asia/Tokyo:20241203T133400
UID:siggraphasia_SIGGRAPH Asia 2024_sess105_papers_423@linklings.com
SUMMARY:PALP: Prompt Aligned Personalization of Text-to-Image Models
DESCRIPTION:Moab Arar (Tel Aviv University), Andrey Voynov and Amir Hertz 
 (Google Research), Omri Avrahami (Hebrew University of Jerusalem), Shlomi 
 Fruchter and Yael Pritch (Google Research), Daniel Cohen-Or (Tel Aviv Univ
 ersity), and Ariel Shamir (Reichman University)\n\nContent creators often 
 aim to create personalized images using personal subjects that go beyond t
 he capabilities of conventional text-to-image models. Additionally, they m
 ay want the resulting image to encompass a specific location, style, ambia
 nce, and more. Existing personalization methods may compromise personaliza
 tion ability or the alignment to complex textual prompts. This trade-off c
 an impede the fulfillment of user prompts and subject fidelity. We propose
  a new approach focusing on personalization methods for a \emph{single} pr
 ompt to address this issue. We term our approach prompt-aligned personaliz
 ation. While this may seem restrictive, our method excels in improving tex
 t alignment, enabling the creation of images with complex and intricate pr
 ompts, which may pose a challenge for current techniques. In particular, o
 ur method keeps the personalized model aligned with a target prompt using 
 an additional score distillation sampling term. We demonstrate the versati
 lity of our method in multi- and single-shot settings and further show tha
 t it can compose multiple subjects or use inspiration from reference image
 s, such as artworks. We compare our approach quantitatively and qualitativ
 ely with existing baselines and state-of-the-art techniques.\n\nRegistrati
 on Category: Full Access, Full Access Supporter\n\nLanguage Format: Englis
 h Language\n\nSession Chair: Kfir Aberman (Decart AI)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_423&sess=sess105
END:VEVENT
END:VCALENDAR