JOURNAL PREPRINT · 2026

Daniel Feijoo, Miguel A. Martínez-Prieto, Anibal Bregon, Álvaro García, Marcos V. Conde

Paper

A survey of efficient deep learning methods for image restoration, covering architectures, training strategies, and practical deployment for real-world restoration pipelines.

JOURNAL PREPRINT · 2026 · NTIRE CHALLENGE

Shuhong Liu, Chenyu Bao, Ziteng Cui, Yun Liu, Xuangeng Chu, Lin Gu, Marcos V. Conde, et al.

Project / arXiv / NTIRE 2026 Challenge / Video

RealX3D is a real-capture 3D benchmark for restoration and reconstruction under in-the-wild degradations — low-light, blur, occlusion, reflection, etc.

CVPR · 2026 · LOVIF WORKSHOP

Daniel Feijoo, Paula Garrido-Mellado, Jaesung Rim, Álvaro García, Marcos V. Conde

GitHub / arXiv / Demo 🤗

DeMoE is the first all-in-one deblurring method that restores images with diverse blur types — global motion, local motion, low-light blur, and defocus.

WACV · 2026

Omar Elezabi, Marcos V. Conde, Zongwei Wu, Radu Timofte

Project / arXiv

A context-aware implicit neural representation that learns professional retouching from before–after pairs and transfers complex Lightroom-style edits to new images with a single example.

arXiv · 2025

Juan C. Benito, Daniel Feijoo, Álvaro García, Marcos V. Conde

arXiv / GitHub

Fast and efficient baselines for real-world low-light image enhancement. Process HD videos at 24 FPS. Fastest model in NTIRE challenges.

ICCV · 2025

Marcos V. Conde, Zihao Lu, Radu Timofte

Paper / GitHub

We propose the first approach that introduces language and explicit control into the image processing and editing pipeline. PixTalk is a vision-language multi-task image processing model, guided using text instructions. Our method is able to perform the most popular techniques in photography.

ICCV · 2025

Tim Seizinger, Florin-Alexandru Vasluianu, Marcos V. Conde, Zongwei Wu, Radu Timofte

arXiv, Code and Models upcoming

The best Bokeh rendering method using neural networks and a novel dataset. The neural model allows to control the apertur from f/16 to f/1.8

ICCV Workshop · 2025

Leo Hoshikawa*, Marcos V. Conde*, Takeshi Ohashi, Atsushi Irie

arXiv

We present a novel analysis on compressing neural fields, with the focus on images. We also introduce Adaptive Neural Images (ANI), an efficient neural representation that enables adaptation to different inference or transmission requirements.

CVPR · 2025

Daniel Feijoo, Juan C. Benito, Alvaro Garcia, Marcos V. Conde

GitHub / arXiv / Demo 🤗

DarkIR In low-light conditions, you have noise and blur in the images, yet, previous methods cannot tackle dark noisy images and dark blurry using a single model. We propose the first approach for all-in-one low-light restoration including illumination, noisy and blur enhancement.

ECCV · 2024

Marcos V. Conde, Gregor Geigle, Radu Timofte

project page / GitHub / arXiv / Video / Twitter X / Demo 🤗

InstructIR takes as input a degraded image and a human-written instruction for how to improve that image. The (single) neural model performs all-in-one image restoration. We achieve state-of-the-art results on several restoration tasks including image denoising, deraining, deblurring, dehazing, and (low-light) image enhancement.

ICIP · 2024

Marcos V. Conde, Andy Bigos, Radu Timofte

arXiv

Implicit Neural Representations (INRs) are a novel paradigm for signal representation that have attracted considerable interest for image compression. We explore the critical yet overlooked limiting factors of INRs, such as computational cost, unstable performance, and robustness

ICIP · 2024

Marcos V. Conde, Florin Vasluianu, Radu Timofte

arXiv

We tackle image restoration directly in the RAW domain (yes, it is tricky and a bit crazy).

ICIP · 2024

Omar Elezabi, Marcos V. Conde, Radu Timofte

arXiv / GitHub

First, we propose a novel module that can be integrated into any neural ISP to capture the global context information from the full RAW images. Second, we propose an efficient and simple neural ISP that utilizes our proposed module.

AAAI · 2024

Marcos V. Conde, Javier Vazquez-Corral , Michael S. Brown, Radu Timofte

project page / arXiv / GitHub & Demo / Video / Easter Egg

NILUTs are neural representations of real 3D LUTs for controllable photo-realistic image enhancement and color manipulation. Moreover, a NILUT can be extended to incorporate multiple styles into a single network with the ability to blend styles implicitly.

WACV · 2024

Marcos V. Conde, Florin Vasluianu, Radu Timofte

arXiv / cvf Proceedings / GitHub

We advance RAW sensor images up-scaling (Super-Resolution). We explore diverse image degradations (e.g. Noise, Blur) to emulate a low-resolution RAW image, and we train a neural network to upsample it.

CVPR Workshop · 2023

Tim Seizinger*, Marcos V. Conde*, Manuel Kolmet, Tom E Bishop, Radu Timofte

paper / cvf Proceedings / GitHub

EBokehNet, an efficient state-of-the-art solution for Bokeh effect transformation and rendering. We can render Bokeh from all-in-focus images, or transform the Bokeh of one lens to the effect of another lens without harming the sharp foreground in the image.

CVPR Workshop · 2023

Eduard Zamfir, Marcos V. Conde, Radu Timofte

paper / cvf Proceedings / GitHub

The paper presents an exhaustive study of baseline methods for real-time image SR. The methods allow 60 FPS and even 120 FPS.

WACV · 2023   · (Oral Presentation, Spotlight)

Marcos V. Conde, Florin Vasluianu, Javier Vazquez-Corral, Radu Timofte

arXiv / cvf Proceedings / GitHub

We propose LPIENet, a lightweight network for perceptual image enhancement, with the focus on deploying it on smartphones. The model was tested for image denoising, deblurring, and HDR correction.

ECCV · 2022

Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte

arXiv / eccv Proceedings / GitHub / Demo (3M runs!)

Super-resolution of compressed images using transformers. We use the Swin Transformer V2, to improve SwinIR for image super-resolution, and in particular, the compressed input scenario. Using this method we can tackle the major issues in training transformer vision models, such as training instability, resolution gaps between pre-training and fine-tuning.

ECCV Workshop · 2022

Marcos V. Conde, Radu Timofte, et al.

arXiv / eccv Proceedings / GitHub

This paper introduces the AIM 2022 Challenge on Reversed Image Signal Processing and RAW Reconstruction. We aim to recover raw sensor images from the corresponding RGBs without metadata and, by doing this, "reverse" the ISP transformation.

AAAI · 2022   · (Oral Presentation, Spotlight)

Marcos V. Conde, Steven McDonagh, Matteo Maggioni , Aleš Leonardis, Eduardo Pérez-Pellitero

project page / GitHub / arXiv / Poster

Hybrid model-based and data-driven approach for modelling ISPs using learnable dictionaries. We explore RAW image reconstruction and improve downstream tasks like RAW Image Denoising via raw data augmentation-synthesis.

CVPR Workshop · 2021

Marcos V. Conde, Kerem Turgutlu

arXiv / cvpr Proceedings / GitHub / Kaggle

We were one of the 1st attempts to use CLIP (Contrastive Language-Image Pre-Training) for training a neural network on a variety of art images and descriptions, being able to learn directly from raw descriptions about images, or if available, curated labels.