PulseAugur
EN
LIVE 14:15:05

MaCo-GAN framework improves image super-resolution with contrastive learning

By PulseAugur Editorial · [2 sources] ·

Researchers have developed MaCo-GAN, a new framework for single image super-resolution that addresses artifact generation in conventional Generative Adversarial Networks (GANs). This novel approach replaces the standard adversarial loss with a supervised contrastive objective, utilizing a dynamic synthesizer to create challenging, perceptually plausible fake images. The MaCo-GAN framework trains the generator to distinguish between on-manifold and off-manifold fakes, leading to improved perception-distortion trade-offs across various benchmarks. AI

IMPACT Introduces a novel contrastive learning approach to improve image super-resolution quality by reducing artifacts.

RANK_REASON The cluster contains a research paper detailing a new method for image super-resolution.

Read on arXiv cs.CV →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

MaCo-GAN framework improves image super-resolution with contrastive learning

COVERAGE [2]

  1. arXiv cs.CV TIER_1 English(EN) · Daeyoung Han, Seongmin Hwang, Moongu Jeon ·

    MaCo-GAN: Manifold-Contrastive Adversarial Learning for Single Image Super-Resolution

    arXiv:2606.05068v1 Announce Type: new Abstract: Conventional Generative Adversarial Networks (GANs) for Single Image Super-Resolution (SISR) often struggle with hallucinated artifacts, largely because standard discriminators evaluate overall image naturalness rather than strict c…

  2. arXiv cs.CV TIER_1 English(EN) · Moongu Jeon ·

    MaCo-GAN: Manifold-Contrastive Adversarial Learning for Single Image Super-Resolution

    Conventional Generative Adversarial Networks (GANs) for Single Image Super-Resolution (SISR) often struggle with hallucinated artifacts, largely because standard discriminators evaluate overall image naturalness rather than strict conditional realism. To address this, we propose …

RELATED ENTITIES

RELATED TOPICS