NeurIPS 2023
PromptIR utilizes prompt-based learning to restore images across various degradation types and levels, achieving state-of-the-art results in denoising, deraining, and dehazing, providing an efficient plugin module for image restoration without prior knowledge of corruptions.

CVPR 2023
we develop a Transformer-based framework for burst-image restoration that leverages multi-scale local and non-local features for improved alignment and feature fusion. Its flexible design allows processing bursts of variable sizes.

CVPR 2023
A new architecture to reconstruct a spatially-precise high-quality image from a burst of low-quality raw images.

TPAMI 2022
A novel feature extraction model that obtains a complementary set of features across multiple spatial scales, while maintaining the original high-resolution features to preserve precise spatial details.

CVPR 2022, Oral
An encoder-decoder Transformer for multi-scale local-global representation learning. Restormer is computationally efficient to handle high-resolution images.

CVPR 2022, Oral, Best Paper Finalist
we develop a novel approach by solely focusing on the effective information exchange between burst frames, such that the degradations get filtered out while the actual scene details are preserved and enhanced.

ACM CSUR 2021
This survey provides a comprehensive overview of the Transformer models in the computer vision discipline, comparing advantages and limitations of popular techniques both in terms of architectural design and their experimental value. Finally, we provide an analysis on open research directions and possible future works.

CVPR 2021
A multi-stage architecture that progressively learns restoration functions for the degraded inputs, thereby breaking down the overall recovery process into more manageable steps.

ECCV 2020
A novel feature extraction model that obtains a complementary set of features across multiple spatial scales, while maintaining the original high-resolution features to preserve precise spatial details.

CVPR 2020, Oral
It is a framework that models camera imaging pipeline in forward and reverse directions and allows us to produce any number of realistic image pairs for denoising both in RAW and sRGB spaces.

CVPR 2019, Oral
A new framework for uncertainty based class imbalance learning.

ICCV 2019
It is the first hybrid loss function that jointly performs classification and clustering in a single formulation.

CVPRw 2019, Oral
The first benchmark dataset for instance segmentation in aerial imagery that combines instance-level object detection and pixel-level segmentation tasks.