通过改进损失权重与归一化方法，有效减少VAE训练中的图像伪影。
VIVAT: Virtuous Improving VAE Training through Artifact Mitigation
- 提出五类常见伪影的分类及成因分析，针对性优化训练策略。
- 在多个基准上实现最佳重建指标（PSNR/SSIM）和更高文本生成质量。
- 无需改动网络结构，适合希望提升VAE性能的研究者和工程师使用。
变分自编码器（VAEs）仍是生成式计算机视觉的核心，但其训练常受图像伪影影响，降低重建与生成质量。本文提出VIVAT，一种系统性缓解KL-VAE训练中常见伪影的方法，无需大幅改变网络结构。我们归纳出五类典型伪影——颜色偏移、网格图案、模糊、角落与滴状伪影，并分析其根本原因。通过调整损失权重、采用合适的填充策略，以及引入空间条件归一化（Spatially Conditional Normalization），显著提升了VAE性能。该方法在多个基准上实现了最先进的图像重建指标（PSNR和SSIM），并显著改善了文本到图像生成质量，表现为更高的CLIP得分。在保持KL-VAE框架简洁性的同时，有效应对实际训练挑战，为研究者和实践者提供了可操作的优化路径。
原文摘要 · Abstract (English)
Variational Autoencoders (VAEs) remain a cornerstone of generative computer vision, yet their training is often plagued by artifacts that degrade reconstruction and generation quality. This paper introduces VIVAT, a systematic approach to mitigating common artifacts in KL-VAE training without requiring radical architectural changes. We present a detailed taxonomy of five prevalent artifacts - color shift, grid patterns, blur, corner and droplet artifacts - and analyze their root causes. Through straightforward modifications, including adjustments to loss weights, padding strategies, and the integration of Spatially Conditional Normalization, we demonstrate significant improvements in VAE performance. Our method achieves state-of-the-art results in image reconstruction metrics (PSNR and SSIM) across multiple benchmarks and enhances text-to-image generation quality, as evidenced by superior CLIP scores. By preserving the simplicity of the KL-VAE framework while addressing its practical challenges, VIVAT offers actionable insights for researchers and practitioners aiming to optimize VAE training.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。