PhD student · University of Barcelona
Imanol G. Estepa
I work on unified self-supervised learning: models that learn useful visual representations while keeping the ability to generate, reconstruct, and reason about images.
Supervised by Prof. Petia Radeva, my research explores representation learning, generative modeling, semantic tokenization, and interpretable latent spaces.
Interactive research pages
Start with the visual versions.
Latest
Recent notes.
Publications
Research trail.
CVPR 2026
Learning from Semantic Dictionaries: Discriminative Codebook Contrastive Learning for Unified Visual Representation and Generation
Imanol G. Estepa, Jesus M. Rodriguez de Vera, Bhalaji Nagarajan, Petia Radeva
WACV 2026
Conjuring Positive Pairs for Efficient Unification of Representation Learning and Image Synthesis
Imanol G. Estepa, Jesus M. Rodriguez de Vera, Bhalaji Nagarajan, Petia Radeva
Pattern Recognition
Precision at Scale: Domain-Specific Datasets On-Demand
Jesus M. Rodriguez de Vera, Imanol G. Estepa, Bhalaji Nagarajan, Petia Radeva
MetaFood Workshop @ CVPR 2024 · Oral
LOFI: LOng-tailed FIne-Grained Network for Food Recognition
Jesus M. Rodriguez de Vera, Imanol G. Estepa, Marc Bolanos, Bhalaji Nagarajan, Petia Radeva
ICCV 2023
All4One: Symbiotic Neighbour Contrastive Learning via Self-Attention and Redundancy Reduction
Imanol G. Estepa, Ignacio Sarasua, Bhalaji Nagarajan, Petia Radeva
MADiMa Workshop @ ACMM 2023 · Oral
Dining on Details: LLM-Guided Expert Networks for Fine-Grained Food Recognition
Jesus M. Rodriguez de Vera, Pablo Villacorta Benito, Imanol G. Estepa, Marc Bolanos, Ignacio Sarasua, Bhalaji Nagarajan, Petia Radeva
VIPriors Workshop @ ICCV 2023
Good Fences Make Good Neighbours
Imanol G. Estepa, Jesus M. Rodriguez de Vera, Bhalaji Nagarajan, Petia Radeva
Academic life
Service and talks.
2025+ CVPR and ICCV reviewer.
2024+ WACV reviewer.
2023+ IEEE Transactions on Multimedia reviewer.
2023 Talk at DLBCN, with public recording and slides.
2023 Master's thesis completed with honours: Analysis and improvement proposal on self-supervised deep learning.