텍스트-이미지 생성 모델 학습을 위한 이미지 캡션: 정밀도와 재현율 분석 및 합성 캡션 활용 가능성 탐구
텍스트-이미지 생성 모델 학습에서 캡션의 정밀도가 재현율보다 이미지 생성 성능에 더 큰 영향을 미치며, Large Vision Language Models (LVLM)을 활용하여 생성한 합성 캡션도 인간의 주석과 유사한 경향을 보인다.
The Impact of Image Caption Precision and Recall on Text-to-Image Generation Model Training
Prioritizing precision over recall in image captions, whether human-annotated or synthetically generated, leads to better performance in training text-to-image generation models, particularly in terms of compositional capabilities.
Subject-driven Text-to-Image Generation via Preference-based Reinforcement Learning (RPO)
This research paper introduces RPO, a novel approach for subject-driven text-to-image generation that leverages a λ-Harmonic reward function and preference-based reinforcement learning to efficiently fine-tune diffusion models, achieving state-of-the-art results in generating images faithful to both reference images and textual prompts.
ReNO: Enhancing One-Step Text-to-Image Models by Optimizing Initial Noise Using Reward Models
ReNO is a novel approach that significantly improves the quality and prompt adherence of one-step text-to-image synthesis models by optimizing the initial latent noise vector based on feedback from multiple human preference reward models.
헬로밈: 디퓨전 모델에 고수준의 사실적인 조건을 포함하기 위한 공간 니팅 어텐션 통합
본 논문에서는 텍스트-투-이미지 기반 모델에 어댑터를 삽입하여 기본 모델의 일반화 능력을 유지하면서 복잡한 다운스트림 작업을 수행할 수 있는 효과적인 방법을 제안합니다.
HelloMeme：拡散モデルに高レベルで忠実度の高い条件を埋め込むための空間ニットアテンションの統合
本稿では、テキストから画像への基盤モデルにアダプターを挿入する効果的な手法を提案する。この手法により、ベースモデルの汎化能力を維持しながら、複雑なダウンストリームタスクを実行することができる。
A Comparative Study of Compositional Generation Abilities in Diffusion and Autoregressive Text-to-Image Models
Diffusion-based text-to-image models outperform autoregressive models in compositional generation tasks, suggesting that the inductive bias of next-token prediction alone is insufficient for complex image generation from text.
LLM 기반 희귀 개념 가이드를 활용한 Diffusion Model의 구성적 생성 능력 향상: 희귀 개념에서 빈번한 개념으로 (R2F)
본 논문에서는 사전 훈련된 Diffusion Model이 희귀 개념을 생성하는 데 어려움을 겪는 문제를 해결하기 위해 LLM(Large Language Model)을 활용하여 희귀 개념을 빈번한 개념으로 변환하여 학습 없이도 이미지 생성 품질을 향상시키는 R2F(Rare-to-Frequent) 프레임워크를 제안합니다.
Rare-to-Frequent (R2F): Enhancing Compositional Image Generation in Diffusion Models Using LLM Guidance for Rare Concepts
Leveraging the semantic knowledge of Large Language Models (LLMs) to guide the diffusion process significantly improves the ability of text-to-image diffusion models to generate images from prompts containing rare or unusual compositions of concepts.
GrounDiT: A Training-Free Spatial Grounding Technique for Text-to-Image Generation Using Diffusion Transformers
GROUNDIT, a novel training-free technique, enhances the spatial accuracy of text-to-image generation using Diffusion Transformers by cultivating and transplanting noisy image patches within specified bounding boxes, leading to more precise object placement compared to previous methods.
© 2024 by Linnk AI