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:20260417T190050Z
LOCATION:West Building\, Rooms 118-120
DTSTART;TZID=America/Los_Angeles:20250813T104500
DTEND;TZID=America/Los_Angeles:20250813T123500
UID:siggraph_SIGGRAPH 2025_sess152@linklings.com
SUMMARY:Deep Image Editing
DESCRIPTION:PartEdit: Fine-Grained Image Editing using Pre-Trained Diffusi
 on Models\n\nWe present PartEdit, a novel diffusion-based system enabling 
 precise, text-based edits of object parts without retraining or manual mas
 ks. Optimizing part-aware tokens generates localized non-binary attention 
 maps to guide seamless edits. Our novel blending strategy delivers high-qu
 ality visual resu...\n\n\nAleksandar Cvejic, Abdelrahman Eldesokey, and Pe
 ter Wonka (King Abdullah University of Science and Technology (KAUST))\n--
 -------------------\nIntrinsicEdit: Precise generative image manipulation 
 in intrinsic space\n\nA generative workflow for precise image editing usin
 g an intrinsic-image latent space. Built on RGB-X diffusion, it enables di
 verse edits—like relighting, color changes, and object manipulation—while 
 preserving identity and ameliorating intrinsic-channel entanglement. All t
 his is done wi...\n\n\nLinjie Lyu (Max-Planck-Institute for Informatics & 
 Saarland Informatics Campus, Adobe Research); Valentin Deschaintre, Yannic
 k Hold-Geoffroy, Milos Hasan, and Jae Shin Yoon (Adobe Research); Thomas L
 eimkuehler and Christian Theobalt (Max-Planck-Institute for Informatics & 
 Saarland Informatics Campus); and Iliyan Georgiev (Adobe Research)\n------
 ---------------\nMonetGPT: Solving Puzzles Enhances MLLMs’ Image Retouchin
 g Skills\n\nMonetGPT explores using multimodal large language models (MLLM
 s) for photo retouching by injecting domain knowledge via visual puzzles. 
 These puzzles help MLLMs understand individual operations,  visual aesthet
 ics, and generate expert plans. Our procedural pipeline enables explainabl
 e edits with det...\n\n\nNiladri Shekhar Dutt (University College London (
 UCL)); Duygu Ceylan (Adobe); and Niloy Mitra (University College London (U
 CL), Adobe)\n---------------------\nDeep Image Editing - Interactive Discu
 ssion\n\nAfter the summary presentations, attendees will participate in an
  interactive discussion. Outside the room will be a series of poster board
 s for authors to gather around with the audience. Authors are invited to b
 ring any material related to their paper that could instigate further conv
 ersation such...\n\n---------------------\nInstance Segmentation of Scene 
 Sketches Using Natural Image Priors\n\nINKi enables instance segmentation 
 for scene sketches by adapting image segmentation models with class-agnost
 ic tuning and depth-based refinement. We introduce a new dataset INK-scene
  with diverse styles and demonstrate layered sketch organization for advan
 ced editing, including inpainting occluded ...\n\n\nMia Tang (Stanford Uni
 versity, Carnegie Mellon University); Yael Vinker (Computer Science and Ar
 tificial Intelligence Laboratory (CSAIL), Massachusetts Institute of Techn
 ology (MIT)); Chuan Yan and Lvmin Zhang (Stanford University); and Maneesh
  Agrawala (Stanford University, Roblox Research)\n---------------------\nC
 ora: Correspondence-aware image editing using few step diffusion\n\nCora i
 s a novel diffusion-based image editing method that achieves complex edits
 , such as object insertion, background changes, and non-rigid transformati
 ons, in only four diffusion steps. By leveraging pixel-wise semantic corre
 spondences between source and target, it preserves key elements of the o..
 .\n\n\nAmirhossein Alimohammadi, Aryan Mikaeili, and Sauradip Nag (Simon F
 raser University); Negar Hassanpour (Huawei Canada); and Andrea Tagliasacc
 hi and Ali Mahdavi-Amiri (Simon Fraser University)\n---------------------\
 n3D-Fixup: Advancing Photo Editing with 3D Priors\n\n3D-Fixup enables real
 istic 3D-aware photo editing by leveraging 3D priors and a novel data pipe
 line that extracts training pairs from real-world videos. Its feed-forward
  architecture supports efficient, high-quality edits involving complex 3D 
 transformations while preserving object identity, outperf...\n\n\nYen-Chi 
 Cheng (University of Illinois Urbana-Champaign, Adobe Research); Krishna K
 umar Singh and Jae Shin Yoon (Adobe Research); Alexander Schwing and Liang
 -Yan Gui (University of Illinois Urbana-Champaign); and Matheus Gadelha, P
 aul Guerrero, and Nanxuan Zhao (Adobe Research)\n---------------------\nBe
  Decisive: Noise-Induced Layouts for Multi-Subject Generation\n\nText-to-i
 mage diffusion models struggle with multi-subject generation due to subjec
 t leakage. Prior methods impose external layouts that conflict with the mo
 del’s prior, harming alignment and natural composition. We introduce a met
 hod that leverages the layout encoded in the initial noise, pro...\n\n\nOm
 er Dahary (Tel Aviv University, Snap Research); Yehonathan Cohen (Tel Aviv
  University); Or Patashnik (Tel Aviv University, Snap Research); Kfir Aber
 man (Snap Research); and Daniel Cohen-Or (Tel Aviv University, Snap Resear
 ch)\n\nInterest Area: Research & Education\n\nRecording: Livestreamed, Not
  Livestreamed, Recorded, Not Recorded\n\nRegistration Category: Full Confe
 rence, Virtual Access, Wednesday\n\nSession Chair: Or Patashnik (Tel Aviv 
 University, Snap Research)
END:VEVENT
END:VCALENDAR