PulseAugur
LIVE 08:58:39
research · [2 sources] ·
0
research

NTIRE 2026 Challenge advances remote sensing infrared image super-resolution

By PulseAugur Editorial · Summary by gemini-2.5-flash-lite from 2 sources

Researchers have developed a dual-branch system for enhancing the resolution of remote sensing infrared images, combining a HAT-L branch with a MambaIRv2-L branch. This approach addresses the unique challenges of thermal imagery, such as weak textures and sensitivity to unstable sharpening, by integrating local and global modeling. The system was presented as a solution for the NTIRE 2026 Infrared Image Super-Resolution Challenge, outperforming individual branches in key metrics. AI

Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →

IMPACT Advances techniques for infrared image super-resolution, potentially improving analysis in remote sensing applications.

RANK_REASON This is a research paper detailing a novel method for a specific AI task and its performance in a challenge.

Read on arXiv cs.CV →

NTIRE 2026 Challenge advances remote sensing infrared image super-resolution

COVERAGE [2]

  1. arXiv cs.CV TIER_1 · Xining Ge, Gengjia Chang, Weijun Yuan, Zhan Li, Zhanglu Chen, Boyang Yao, Yihang Chen, Yifan Deng, Shuhong Liu ·

    Dual-Branch Remote Sensing Infrared Image Super-Resolution

    arXiv:2604.10112v2 Announce Type: replace Abstract: Remote sensing infrared image super-resolution aims to recover sharper thermal observations from low-resolution inputs while preserving target contours, scene layout, and radiometric stability. Unlike visible-image super-resolut…

  2. arXiv cs.CV TIER_1 · Adrien Gressin ·

    The First Challenge on Remote Sensing Infrared Image Super-Resolution at NTIRE 2026: Benchmark Results and Method Overview

    This paper presents the NTIRE 2026 Remote Sensing Infrared Image Super-Resolution (x4) Challenge, one of the associated challenges of NTIRE 2026. The challenge aims to recover high-resolution (HR) infrared images from low-resolution (LR) inputs generated through bicubic downsampl…

RELATED ENTITIES

RELATED TOPICS