Skip to main content
Johannes Roth
  • experience
  • work
  • contact

Projects

Research tools, publications, and software

01

Datasets & Research Tools

Dataset

ReLAION-2B Natural

Naturalness scores for 2.1B images from ReLAION-2B, identifying ~500M photographic images suitable for vision research.

167GB • CLIP ViT-H/14
Library

thingsvision

Feature extraction for 50+ vision models with a unified API. CLIP, ViT, ResNet, and more.

Contributor • Python
Tool

cvmanova_python

Cross-validated MANOVA for multivariate pattern analysis in fMRI.

Python
Tool

pysearchlight

Customizable searchlight analysis for fMRI data.

Python
02

Papers & Conferences

CCN 2025 · Oral

How to sample the world for understanding the visual system

Roth & Hebart
Comms Psych 2025

Ten principles for reliable, efficient, and adaptable coding

Roth et al.
J Neurosci 2025

Fine-grained image and category information in ventral visual pathway

Badwal, Bergmann, Roth et al.
VSS 2023 · Poster

High stimulus presentation rates for fMRI

Roth, Miyawaki & Hebart
Cosyne 2021 · Poster

Preferred stimuli for individual voxels in the human visual system

Roth, Seeliger, Schmid & Hebart
MICCAI BraTS 2021

Multi-plane UNet++ Ensemble for Glioblastoma Segmentation

Roth, Keller et al.
• 2025 CMBB Replication Award - for reliable coding practices in neuroscience
© 2026 Johannes Roth
Email GitHub LinkedIn Hugging Face Scholar