Publications by Majed El Helou

BIGPrior: Towards Decoupling Learned Prior Hallucination and Data Fidelity in Image Restoration

Classic image-restoration algorithms use a variety of priors, either implicitly or explicitly. Their priors are hand-designed and their corresponding weights are heuristically assigned. Hence, deep learning methods often produce superior image restoration ...

2022

PoGaIN: Poisson-Gaussian Image Noise Modeling From Paired Samples

Sabine Süsstrunk, Majed El Helou, Kaan Okumus

Image noise can often be accurately fitted to a Poisson-Gaussian distribution. However, estimating the distribution parameters from a noisy image only is a challenging task. Here, we study the case when paired noisy and noise-free samples are accessible. N ...

IEEE Institute of Electrical and Electronics Engineers2022

Image Denoising with Control over Deep Network Hallucination

Sabine Süsstrunk, Majed El Helou, Qiyuan Liang

Deep image denoisers achieve state-of-the-art results but with a hidden cost. As witnessed in recent literature, these deep networks are capable of overfitting their training distributions, causing inaccurate hallucinations to be added to the output and ge ...

Society for Imaging Science and Technology (IS&T)2022

Adaptive extreme video completion

Sabine Süsstrunk, Radhakrishna Achanta, Ruofan Zhou, Majed El Helou

A computer-implemented video completion method is proposed for reconstructing a sparsely sampled video comprising sparsely sampled video frames comprising sparse sets of picture element intensity values. The method comprises: distributing the sparse sets o ...

2021

Fidelity Estimation Improves Noisy-Image Classification with Pretrained Networks

Sabine Süsstrunk, Majed El Helou, Deblina Bhattacharjee, Xiaoyu Lin

Image classification has significantly improved using deep learning. This is mainly due to convolutional neural networks (CNNs) that are capable of learning rich feature extractors from large datasets. However, most deep learning classification methods are ...

IEEE Institute of Electrical and Electronics Engineers2021

Deep Gaussian denoiser epistemic uncertainty and decoupled dual-attention fusion

Sabine Süsstrunk, Majed El Helou, Xiaoyu Lin, Xiaoqi Ma

Following the performance breakthrough of denoising networks, improvements have come chiefly through novel architecture designs and increased depth. While novel denoising networks were designed for real images coming from different distributions, or for sp ...

2021

Deep Image Restoration: Between Data Fidelity and Learned Priors

Majed El Helou

Image restoration reconstructs, as faithfully as possible, an original image from a potentially degraded version of it. Image degradations can be of various types, for instance haze, unwanted reflections, optical or spectral aberrations, or other physicall ...

EPFL2021

NTIRE 2021 Depth Guided Image Relighting Challenge

Sabine Süsstrunk, Radu Timofte, Ruofan Zhou, Majed El Helou

Image relighting is attracting increasing interest due to its various applications. From a research perspective, image relighting can be exploited to conduct both image normalization for domain adaptation, and also for data augmentation. It also has multip ...

IEEE COMPUTER SOC2021

Stochastic Frequency Masking to Improve Super-Resolution and Denoising Networks

Sabine Süsstrunk, Ruofan Zhou, Majed El Helou

Super-resolution and denoising are ill-posed yet fundamental image restoration tasks. In blind settings, the degradation kernel or the noise level are unknown. This makes restoration even more challenging, notably for learning-based methods, as they tend t ...

2020

Blind Universal Bayesian Image Denoising With Gaussian Noise Level Learning

Sabine Süsstrunk, Majed El Helou

Blind and universal image denoising consists of using a unique model that denoises images with any level of noise. It is especially practical as noise levels do not need to be known when the model is developed or at test time. We propose a theoretically-gr ...

IEEE2020

W2S: Microscopy Data with Joint Denoising and Super-Resolution for Widefield to SIM Mapping

Daniel Sage, Sabine Süsstrunk, Thierry Laroche, Arne Seitz, Ruofan Zhou, Majed El Helou

In fluorescence microscopy live-cell imaging, there is a critical trade-off between the signal-to-noise ratio and spatial resolution on one side, and the integrity of the biological sample on the other side. To obtain clean high-resolution (HR) images, one ...

2020

Realizability of planar point embeddings from angle measurements

Frederike Dümbgen, Adam James Scholefield, Majed El Helou

Localization of a set of nodes is an important and a thoroughly researched problem in robotics and sensor networks. This paper is concerned with the theory of localization from inner-angle measurements. We focus on the challenging case where no anchor loca ...

IEEE2020

Divergence-Based Adaptive Extreme Video Completion

Sabine Süsstrunk, Ruofan Zhou, Fabrice Jean Guibert, Frank Schmutz, Majed El Helou

Extreme image or video completion, where, for instance, we only retain 1% of pixels in random locations, allows for very cheap sampling in terms of the required pre-processing. The consequence is, however, a reconstruction that is challenging for humans an ...

IEEE2020

AIM 2020: Scene Relighting and Illumination Estimation Challenge

Sabine Süsstrunk, Radu Timofte, Matthew Brown, Ruofan Zhou, Majed El Helou, Tongtong Zhao, Shanshan Zhao, Yu Zhu, Xin Huang, Yuzhong Liu

We review the AIM 2020 challenge on virtual image relighting and illumination estimation. This paper presents the novel VIDIT dataset used in the challenge and the different proposed solutions and final evaluation results over the 3 challenge tracks. The f ...

2020

AL2: Progressive Activation Loss for Learning General Representations in Classification Neural Networks

Sabine Süsstrunk, Frederike Dümbgen, Majed El Helou

The large capacity of neural networks enables them to learn complex functions. To avoid overfitting, networks however require a lot of training data that can be expensive and time-consuming to collect. A common practical approach to attenuate overfitting i ...

IEEE2020

Blind Universal Bayesian Image Denoising with Gaussian Noise Level Learning

Sabine Süsstrunk, Majed El Helou

Blind and universal image denoising consists of a unique model that denoises images with any level of noise. It is especially practical as noise levels do not need to be known when the model is developed or at test time. We propose a theoretically-grounded ...

2019

Solving the Depth Ambiguity in Single-Perspective Images

Sabine Süsstrunk, Marjan Shahpaski, Majed El Helou

Scene depth estimation is gaining in importance as more and more AR/VR and robot vision applications are developed. Conventional depth-from-defocus techniques can passively provide depth maps from a single image. This is especially advantageous for moving ...

2019

A comparative study on wavelets and residuals in deep super resolution

Sabine Süsstrunk, Ruofan Zhou, Fayez Lahoud, Majed El Helou

Despite the advances in single-image super resolution using deep convolutional networks, the main problem remains unsolved: recovering fine texture details. Recent works in super resolution aim at modifying the training of neural networks to enable the rec ...

2019

Drone Shadow Tracking

Sabine Süsstrunk, Ruofan Zhou, Majed El Helou

Aerial videos taken by a drone not too far above the surface may contain the drone's shadow projected on the scene. This deteriorates the aesthetic quality of videos. With the presence of other shadows, shadow removal cannot be directly applied, and the sh ...

2019

AIM 2019 Challenge on Image Extreme Super-Resolution: Methods and Results

Radu Timofte, Xin Yang, Jia Yu, Ruofan Zhou, Muhammad Haris, Majed El Helou, Peng Yi, Jiayi Ma, Wei Li, Hanwen Liu, Liying Lu

This paper reviews the AIM 2019 challenge on extreme image super-resolution, the problem of restoring of rich details in a low resolution image. Compared to previous, this challenge focuses on an extreme upscaling factor, x16, and employs the novel DIVerse ...

IEEE COMPUTER SOC2019