PortraitGen: Exemplar-Driven GRPO with Dual-Reward Guidance for Photorealistic Portrait Generation

• URL: http://arxiv.org/abs/2606.26930v1
• Date: Thu, 25 Jun 2026 12:05:40 GMT
• Title: PortraitGen: Exemplar-Driven GRPO with Dual-Reward Guidance for Photorealistic Portrait Generation
• Abstract summary: We propose PortraitGen, a novel framework tailored for photorealistic portrait generation.<n>First, we break inherent generative boundaries by directly introducing real images into the GRPO sampling groups.<n>We introduce a complementary dual-reward mechanism: OmniReward for general quality and AI-Portrait for human-centric fidelity.
• Score: 54.85864545266293
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Reinforcement Learning like Group Relative Policy Optimization (GRPO) has significantly advanced text-to-image post-training. However, current methods often favor superficial aesthetics, such as over-saturated colors, leaving critical flaws like AI artifacts and biological implausibilities unresolved. We attribute these limitations to two primary factors: (1) The absence of real images during post-training confines GRPO sampling to the original distribution, failing to break inherent generative boundaries; (2) the optimization process lacks specific rewards targeting fine-grained artifacts like overly oily skin and other AI artifacts. To address this, we propose PortraitGen, a novel framework tailored for photorealistic portrait generation. First, we break inherent generative boundaries by directly introducing real images into the GRPO sampling groups, where image inversion is employed to obtain their transition probabilities and latents. Second, to explicitly steer the model toward photorealism, we introduce a complementary dual-reward mechanism: OmniReward for general quality and AI-Portrait for human-centric fidelity. Furthermore, we curate PortraitBench, a comprehensive portrait-centric benchmark. Extensive experiments demonstrate that PortraitGen significantly outperforms existing baselines, effectively suppressing AI artifacts and achieving unprecedented photorealism.

Related papers

• Detecting AI-Generated Images via Distributional Deviations from Real Images [6.615773227400183]
  We propose a Masking-based Pre-trained model Fine-Tuning (MPFT) strategy, which introduces a Texture-Aware Masking (TAM) mechanism to mask textured areas containing generative model-specific patterns during fine-tuning.<n>Our method, fine-tuned with only a minimal number of images, significantly outperforms existing approaches, achieving up to 98.2% and 94.6% average accuracy on the two datasets, respectively.
  arXiv  Detail & Related papers  (2026-01-07T05:00:13Z)
• RealGen: Photorealistic Text-to-Image Generation via Detector-Guided Rewards [53.25632969696776]
  We propose RealGen, a text-to-image framework for photorealistic image generation.<n>Inspired by adversarial generation, RealGen introduces a "Detector Reward" mechanism, which quantifies artifacts and assesses realism.<n>Experiments demonstrate that RealGen significantly outperforms general models like GPT-Image-1 and Qwen-Image, as well as specialized photorealistic models like FLUX-Krea.
  arXiv  Detail & Related papers  (2025-11-29T12:52:26Z)
• Timestep-Aware Diffusion Model for Extreme Image Rescaling [47.89362819768323]
  We propose a novel framework called Timestep-Aware Diffusion Model (TADM) for extreme image rescaling.<n>TADM performs rescaling operations in the latent space of a pre-trained autoencoder.<n>It effectively leverages powerful natural image priors learned by a pre-trained text-to-image diffusion model.
  arXiv  Detail & Related papers  (2024-08-17T09:51:42Z)
• Boost Your Human Image Generation Model via Direct Preference Optimization [5.9726297901501475]
  Human image generation is a key focus in image synthesis due to its broad applications, but even slight inaccuracies in anatomy, pose, or details can compromise realism.<n>We explore Direct Preference Optimization (DPO), which trains models to generate preferred (winning) images while diverging from non-preferred (losing) ones.<n>We propose an enhanced DPO approach that incorporates high-quality real images as winning images, encouraging outputs to resemble real images rather than generated ones.
  arXiv  Detail & Related papers  (2024-05-30T16:18:05Z)
• Self-Adaptive Reality-Guided Diffusion for Artifact-Free Super-Resolution [47.29558685384506]
  Artifact-free super-resolution (SR) aims to translate low-resolution images into their high-resolution counterparts with a strict integrity of the original content. Traditional diffusion-based SR techniques are prone to artifact introduction during iterative procedures. We propose Self-Adaptive Reality-Guided Diffusion to identify and mitigate the propagation of artifacts.
  arXiv  Detail & Related papers  (2024-03-25T11:29:19Z)
• HyperReenact: One-Shot Reenactment via Jointly Learning to Refine and Retarget Faces [47.27033282706179]
  We present our method for neural face reenactment, called HyperReenact, that aims to generate realistic talking head images of a source identity. Our method operates under the one-shot setting (i.e., using a single source frame) and allows for cross-subject reenactment, without requiring subject-specific fine-tuning. We compare our method both quantitatively and qualitatively against several state-of-the-art techniques on the standard benchmarks of VoxCeleb1 and VoxCeleb2.
  arXiv  Detail & Related papers  (2023-07-20T11:59:42Z)
• CtlGAN: Few-shot Artistic Portraits Generation with Contrastive Transfer Learning [77.27821665339492]
  CtlGAN is a new few-shot artistic portraits generation model with a novel contrastive transfer learning strategy. We adapt a pretrained StyleGAN in the source domain to a target artistic domain with no more than 10 artistic faces. We propose a new encoder which embeds real faces into Z+ space and proposes a dual-path training strategy to better cope with the adapted decoder.
  arXiv  Detail & Related papers  (2022-03-16T13:28:17Z)
• PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models [77.32079593577821]
  PULSE (Photo Upsampling via Latent Space Exploration) generates high-resolution, realistic images at resolutions previously unseen in the literature. Our method outperforms state-of-the-art methods in perceptual quality at higher resolutions and scale factors than previously possible.
  arXiv  Detail & Related papers  (2020-03-08T16:44:31Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.