Iliyan Georgiev

Light transporter. Opinions are random; expect bias. Adobe Research, previously Autodesk, Solid Angle.

[Iliyan Georgiev @iliyang]

Got a paper around appearance at CVPR 2026? Make the most of it at the AppX workshop! Submit by April 7:
https://t.co/ZFZp34mLY7

@ele_g2 @vdeschaintre @michi_fischer
@giuvecchio95

#CVPR2026 #ComputerVision #NeuralRendering #AppearanceUnderstanding #AppX2026

[Valentin Deschaintre @vdeschaintre]

I was at #AdobeMAX this week to present #projectSurfaceSwap! Our surface selection and replacement techs!
Check it out here: https://t.co/l76025JQrb

What a blast the Sneaks were!

[Adobe Research @AdobeResearch]

Adobe Research is heading to #SIGGRAPH2025, the premiere conference for computer graphics and interactive techniques!
From generative image editing to 3D reconstruction, we’re sharing 25 papers this year – this teaser is just a glimpse. 🔗 https://t.co/FOvDUgbet6
Come find us at @siggraph and explore the future of creativity!

[Iliyan Georgiev @iliyang]

With Linjie Lyu, @vdeschaintre, @yannickhold, @MilosHasan, @JaeShinYoon2, Thomas Leimkühler, Christian Theobald

Code will be released shortly.

[Iliyan Georgiev @iliyang]

Editing appearance, geometry, lighting with precision is easy with a 3D scene representation. But it's so much more difficult with just an image or photograph.

Enter IntrinsicEdit: Precise generative image manipulation in intrinsic space (@siggraph 2025)!
https://t.co/ebO7f6iEgG

[Iliyan Georgiev @iliyang]

Without no task-specific training, we show state-of-the-art results on material editing, object removal & insertion, relighting.

Our method can preserve identity of unedited image aspects, and thanks to editing in intrinsic-image space, it provides unprecedented control.

[Zheng Zeng @zheng95z]

I didn't even expect it will work on anime images! Check this out!

[Zheng Zeng @zheng95z]

Cool! Try the rgbx wrapper in ComfyUI made by @toyxyz3. If you're interested about the technical details, check out our paper at https://t.co/t74kPN3FyK.

[toyxyz @toyxyz3]

I created the ComfyUI_rgbx_Wrapper node. Try using rgbx in ComfyUI! #AIイラスト #AI #Comfyui

[Valentin Deschaintre @vdeschaintre]

🎓 We introduce SAMa! A material selection and segmentation model on 3D models in any format (3DGS, NeRF, Mesh).
Given a user click, we propose to select all regions on an objects with the same material. We can also do segmentation in under a minute: https://t.co/cl3KvzBLcw

[Zheng Zeng @zheng95z]

We’re thrilled to release the implementation and model weights for our SIGGRAPH 2024 paper, "RGB↔X: Image Decomposition and Synthesis Using Material- and Lighting-aware Diffusion Models"!

📂 GitHub Repository: https://t.co/xcpDTskBTn

RGB↔X is a unified diffusion-based framework that enables realistic image analysis (intrinsic channel estimation, referred to as RGB→X) and synthesis (realistic rendering from intrinsic channels, referred to as X→RGB).

This framework explores the connections between diffusion models, realistic rendering, and intrinsic decomposition. We believe this can benefit a wide range of downstream tasks, including material editing, relighting, and realistic rendering from simple or under-specified scene definitions.

TL;DR for graphics people: You can now use diffusion models to eliminate or simulate global illumination in image-space on G-buffers.

🌐 Project Page: https://t.co/j8frbAAn2r

[Joey Litalien @JoeyLitalien]

We present a learning-based rendering method that uses normalizing flows to efficiently importance sample product distributions comprising distant emitters (i.e., HDRI lighting). This is joint work with my amazing collaborators @MilosHasan, @fujun_luan, @krishnamullia, and @iliyang at Adobe Research. 1/5

[Iliyan Georgiev @iliyang]

@bai_liping @_bingxu @EGSympRendering @tzumaoli That's before taxes

[Bing Xu @_bingxu]

What a pleasant big surprise. Free to share a $1M secret to get a best paper award: to get Iliyan presenting the paper. @EGSympRendering @iliyang. Congrats to @tzumaoli, Trevor and Ravi.

[Iliyan Georgiev @iliyang]

No more huge output PDFs in Overleaf. Put this file in your project root and it will compress non-JPEG images. Default quality settings are high, works even for noisy images.
https://t.co/Mj5PlWGqAR

[Adobe Research @AdobeResearch]

Check out Adobe Research's new work in Computer Vision at #CVPR2024! https://t.co/hXaFfCbhmX

[Élie Michel 🍪 @exppad]

Our work on #NBVH (Neural ray queries with Bounding Volume Hierarchies) will be presented this summer at #SIGGRAPH2024!
What's that? See🧵below ⤵️

📃Source Code: https://t.co/ZR6Ji7Hlz1
🌐Project Page/Paper: https://t.co/ooiiL2PHX0

(ft. @PhilippeWeier @iliyang @boubek et al.)

[Weier Philippe @PhilippeWeier]

Ever wanted to load a large asset in your GPU path tracer, only to find it doesn't fit in memory? Then check out our #SIGGRAPH2024 paper “N-BVH: Neural ray queries with bounding volume hierarchies” (with Alexander Rath, @exppad, @iliyang, Philipp Slusallek and @boubek)

[Valentin Deschaintre @vdeschaintre]

📸You took a room photo or generated one with your favourite Diffusion model? Want to turn it into 3D but find finding and placing the assets is boring?

📚Checkout PSDR-Room for turning a single photo to scene using differentiable rendering & retrieval! https://t.co/CW8505vTSv