Alexandros Graikos

Selected research

On the spectral properties of generative denoiser Jacobians

Alexandros Graikos, Nebojsa Jojic, Dimitris Samaras

Preprint

In denoisers, the eigenvectors of the Jacobian encode generation quality. Increasing the spectrum during training improves generation quality.

Code / arXiv

Fast constrained sampling in pre-trained diffusion models

Alexandros Graikos, Nebojsa Jojic, Dimitris Samaras

NeurIPS 2025

Speeding up training-free inference with a simple approximation. We made a game that uses this algorithm to generate images with constraints in real-time [Gameplay 1 / Gameplay 2].

Code / arXiv

PixCell: A generative foundation model for digital histopathology images

Alexandros Graikos*, Srikar Yellapragada*, Zilinghan Li, Kostas Triaridis, Varun Belagali, Tarak Nath Nandi, Karen Bai, Beatrice S. Knudsen, Tahsin Kurc, Rajarsi R. Gupta, Prateek Prasanna, Ravi K Madduri, Joel Saltz, Dimitris Samaras

Preprint

The first generative foundation model for digital histopathology images.

PixCell-1024 🤗 / PixCell-256 🤗 / arXiv

ZoomLDM: Latent Diffusion Model for multi-scale image generation

Alexandros Graikos*, Srikar Yellapragada*, Kostas Triaridis, Prateek Prasanna, Rajarsi R. Gupta, Joel Saltz, Dimitris Samaras

CVPR 2025

A multi-scale diffusion model to generate very large coherent histopathology images.

Code / arXiv / Large Image Viewer

Learned representation-guided diffusion models for large-image generation

Alexandros Graikos*, Srikar Yellapragada*, Minh-Quan Le, Saarthak Kapse, Prateek Prasanna, Joel Saltz, Dimitris Samaras

CVPR 2024

Self-supervised models provide patch annotations for diffusion models. We extrapolate from patches to synthesize large images.

Code / arXiv

Diffusion models as plug-and-play priors

Alexandros Graikos, Esmeralda S. Whitammer, Nebojsa Jojic, Dimitris Samaras

NeurIPS 2022

Any inference task can use a diffusion model as an image prior. You can even solve the traveling salesman problem!

Code / arXiv

Recent Publications

2026

On the spectral properties of generative denoiser Jacobians

Alexandros Graikos, Nebojsa Jojic, Dimitris Samaras,
Preprint

BigCell: Generating Gigapixel Whole-Slide Images

Srikar Yellapragada, Alexandros Graikos, Zilinghan Li, Kostas Triaridis, Tarak Nath Nandi, Ji Dong K Bai, Beatrice Knudsen, Tahsin Kurc, Prateek Prasanna, Rajarsi R. Gupta, Ravi K Madduri, Joel Saltz, Dimitris Samaras
NeurIPS 2026

Eyes of the beholder: attention-guided metamers reveal divergent scene representations

Ritik Raina, Abe Leite, Alexandros Graikos, Seoyoung Ahn, Gregory J Zelinsky
CCN 2026

Poppy: Polarization-based Plug-and-Play Guidance for Enhancing Monocular Normal Estimation

Irene Kim, Sai Tanmay Reddy Chakkera, Alexandros Graikos, Dimitris Samaras, Akshat Dave
ECCV 2026 · Best paper honorable mention

Gen-SIS: Generative Self-augmentation Improves Self-supervised Learning

Varun Belagali*, Srikar Yellapragada*, Alexandros Graikos, Saarthak Kapse, Zilinghan Li, Tarak Nath Nandi, Ravi K Madduri, Prateek Prasanna, Joel Saltz, Dimitris Samaras
DataCV Workshop @ CVPR 2026 · Best paper award

GriDiT: Factorized Grid-Based Diffusion for Efficient Long Image Sequence Generation

Snehal Singh Tomar, Alexandros Graikos, Arjun Krishna, Dimitris Samaras, Klaus Mueller
TMLR 2026

Structured image representation learning for flow-matching models

Alexandros Graikos, Kostas Triaridis, Esmeralda S. Whitammer, Dimitris Samaras
ICLR 2026 2nd Workshop on Deep Generative Model in Machine Learning: Theory, Principle and Efficacy

Generating metamers of human scene understanding

Ritik Raina, Abe Leite, Alexandros Graikos, Seoyoung Ahn, Dimitris Samaras, Greg Zelinsky
ICLR 2026

Selected research

On the spectral properties of generative denoiser Jacobians

Fast constrained sampling in pre-trained diffusion models

PixCell: A generative foundation model for digital histopathology images

ZoomLDM: Latent Diffusion Model for multi-scale image generation

Learned representation-guided diffusion models for large-image generation

Diffusion models as plug-and-play priors

Recent Publications

2026

On the spectral properties of generative denoiser Jacobians

BigCell: Generating Gigapixel Whole-Slide Images

Eyes of the beholder: attention-guided metamers reveal divergent scene representations

Poppy: Polarization-based Plug-and-Play Guidance for Enhancing Monocular Normal Estimation

Gen-SIS: Generative Self-augmentation Improves Self-supervised Learning

GriDiT: Factorized Grid-Based Diffusion for Efficient Long Image Sequence Generation

Structured image representation learning for flow-matching models

Generating metamers of human scene understanding

2025

Fast constrained sampling in pre-trained diffusion models

PixCell: A generative foundation model for digital histopathology images

Mitigating Diffusion Model Hallucinations with Dynamic Guidance

Pathology Image Compression with Pre-trained Autoencoders

PathSegDiff: Pathology Segmentation using Diffusion model representations

Seen2Scene: a generative model of fixation-by-fixation scene understanding

ZoomLDM: Latent Diffusion Model for multi-scale image generation