Mohammadreza Salehi

About Me

I received my PhD from the University of Amsterdam, where I was a member of the QUVA Lab and the VIS Lab, supervised by Yuki Asano, Cees Snoek, and Efstratios Gavves. I was also part of the ELLIS PhD Program, in collaboration with Qualcomm AI Research.

My research focuses on pretraining and post-training vision foundation models. On the pretraining side, I contributed to Franca, the first fully open-source vision foundation model (open data, code, and weights) to be competitive with DINOv2, and developed SIGMA, a masked video modeling method that replaces pixel reconstruction with semantic feature targets learned through Sinkhorn clustering. On the post-training side, I developed methods including TimeTuning, MoSiC, NeCo, and 3DPoV, which improve the scene and 3D understanding of pretrained models by further training them on videos or images.

During my internships, I worked on efficient causal video editing models built on pretrained image generators such as Stable Diffusion 1.5, Stable Diffusion 3.5, and FLUX. This work resulted in a new family of diffusion models, RFDM (Residual Flow Diffusion Models). My research also spans machine learning safety, particularly outlier detection, which helps AI systems reliably identify and handle unexpected inputs.

News

• [Aug 2026] Our paper Crane, on zero-shot anomaly detection, is accepted at BMVC'26!
• [Jun 2026] Presented a poster at NCCV'26 in Doorwerth.
• [May 2026] Our paper 3DPoV, on improving 3D understanding via patch ordering on videos, is accepted at ICML'26!
• [Feb 2026] 2 papers accepted at CVPR'26: RFDM and Franca!
• [Feb 2026] Our work RFDM, efficient causal video editing from my Samsung internship, is on arXiv. Check it out!
• [Sep 2025] 1 paper accepted at UniReps workshop at NeurIPS'25!
• [Sep 2025] 1 paper accepted at NeurIPS'25!
• [Aug 2025] Presented our work MoSiC at Oxford VGG group
• [Jul 2025] 1 paper accepted at ICCV'25!
• [Jun 2025] Started my internship at GenAI group, Samsung AI Center Cambridge.
• [May 2025] 1 paper accepted at SynData4CV workshop at CVPR'25!
• [May 2025] Teaching assistant for Oxford Machine Learning summer school!
• [Jan 2025] Our paper, NeCo, is accepted at ICLR'25!
• [Sep 2024] 1 paper accepted at ACCV'24 as Oral!
• [Aug 2024] Our work NeCo is on arxiv. Check it out!
• [Jul 2024] 3 papers accepted at ECCV'24!
• [Apr 2024] We are holding a tutorial on Time is precious: Self-Supervised Learning Beyond Images at ECCV'24.
• [Mar 2024] Attended the winter school on foundation models in Amsterdam.
• [Mar 2024] Gave a talk at 9th Winter Seminar Series in advanced topics of Computer Science and Engineering at Sharif Univeristy of Technology.
• [Feb 2024] Serving as a reviewer for ECCV'24.

Publications

Crane: Context-Guided Prompt Learning and Attention Refinement for Zero-Shot Anomaly Detection

Alireza Salehi, Mohammadreza Salehi, Reshad Hosseini, Cees G. M. Snoek, Makoto Yamada, Mohammad Sabokrou

BMVC, 2026

PDF CODE

3DPoV: Improving 3D Understanding via Patch Ordering on Videos

Ioana Simion*, Mohammadreza Salehi*, Shashanka Venkataramanan, Cees G. M. Snoek, Yuki M. Asano

ICML, 2026

PDF CODE

RFDM: Residual Flow Diffusion Model for Efficient Causal Video Editing

Mohammadreza Salehi*, Mehdi Noroozi*, Luca Morreale, Ruchika Chavhan, Malcolm Chadwick, Alberto Gil Ramos, Abhinav Mehrotra

CVPR, 2026

PDF PROJECT

Franca: Nested Matryoshka Clustering for Scalable Visual Representation Learning

Shashanka Venkataramanan, Valentinos Pariza, Mohammadreza Salehi, Lukas Knobel, Spyros Gidaris, Elias Ramzi, Andrei Bursuc, Yuki M. Asano

CVPR, 2026

PDF CODE

DISCOVR: Self-supervised Learning of Echocardiographic Video Representations via Online Cluster Distillation

Divyanshu Mishra, Mohammadreza Salehi, Pramit Saha, Olga Patey, Aris T. Papageorghiou, Yuki M. Asano, J. Alison Noble

NeurIPS, 2025

PDF CODE

MoSiC: Optimal-Transport Motion Trajectory for Dense Self-Supervised Learning

Mohammadreza Salehi*, Shashanka Venkataramanan*, Ioana Simion, Efstratios Gavves, Cees G. M. Snoek, Yuki M Asano

ICCV, 2025

PDF CODE

NeCo: Improving DINOv2's spatial representations in 19 GPU hours with Patch Neighbor Consistency

Valentinos Pariza*, Mohammadreza Salehi*, Gertjan J. Burghouts, Francesco Locatello, Yuki M Asano

ICLR, 2025

PDF CODE

SIGMA: Sinkhorn-Guided Masked Video Modeling

Mohammadreza Salehi*, Michael Dorkenwald*, Fida Mohammad Thoker*, Efstratios Gavves, Cees G. M. Snoek, Yuki M. Asano

ECCV, 2024

PDF CODE

Redefining Normal: A Novel Object-Level Approach for Multi-Object Novelty Detection

Mohammadreza Salehi, Nikolaos Apostolikas, Efstratios Gavves, Cees G. M. Snoek, Yuki M. Asano

ACCV, 2024 (Oral)

PDF CODE

GeneralAD: Anomaly Detection Across Domains by Attending to Distorted Features

Luc P.J. Sträter*, Mohammadreza Salehi*, Efstratios Gavves, Cees G. M. Snoek, Yuki M. Asano

ECCV, 2024

PDF CODE

Time Does Tell: Self-Supervised Time-Tuning of Dense Image Representations

Mohammadreza Salehi, Efstratios Gavves, Cees G. M. Snoek, and Yuki M. Asano

ICCV, 2023

PDF CODE

Fake It Until You Make It : Towards Accurate Near-Distribution Novelty Detection

Hossein Mirzaei, Mohammadreza Salehi, Sajjad Shahabi, Efstratios Gavves, Cees G. M. Snoek, Mohammad Sabokrou, Mohammad Hossein Rohban

ICLR, 2023

PDF CODE

Forecasting influenza hemagglutinin mutations through the lens of anomaly detection

Ali Garjani, Atoosa Malemir Chegini, Mohammadreza Salehi, Alireza Tabibzadeh, Parastoo Yousefi, Mohammad Hossein Razizadeh, Moein Esghaei, Maryam Esghaei, Mohammad Hossein Rohban

Scientific Reports, 2023

PDF CODE

A Unified Survey on Anomaly, Novelty, Open-Set, and Out-of-Distribution Detection: Solutions and Future Challenges

Mohammadreza Salehi, Hossein Mirzaei, Dan Hendrycks, Yixuan Li, Mohammad Hossein Rohban, Mohammad Sabokrou

TMLR, 2022

PDF CODE

Multiresolution Knowledge Distillation for Anomaly Detection

Mohammadreza Salehi, Niousha Sadjadi, Soroosh Baselizadeh, Mohammad Hossein Rohban, Hamid R. Rabiee

CVPR, 2021

PDF CODE

Teaching

• [Apr 2025] TA for UvA foundation models course
• [Nov 2024] Head TA for UvA deep learning 1 course
• [Apr 2024] TA for UvA foundation models course
• [Nov 2023] Head TA for UvA deep learning 1 course

Contact