Recent studies have explored using pretrained Vision Foundation Models (VFMs) such as DINO for generative autoencoders, showing strong generative performance. Unfortunately, existing approaches often suffer from limited reconstruction fidelity due to the loss of high-frequency details. In this work, we present the DINO Spherical Autoencoder (DINO-SAE), a framework that bridges semantic representation and pixel-level reconstruction. Our key insight is that semantic information in contrastive representations is primarily encoded in the direction of feature vectors, while forcing strict magnitude matching can hinder the encoder from preserving fine-grained details. To address this, we introduce Hierarchical Convolutional Patch Embedding module that enhances local structure and texture preservation, and Cosine Similarity Alignment objective that enforces semantic consistency while allowing flexible feature magnitudes for detail retention. Furthermore, leveraging the observation that SSL-based foundation model representations intrinsically lie on a hypersphere, we employ Riemannian Flow Matching to train a Diffusion Transformer (DiT) directly on this spherical latent manifold. Experiments on ImageNet-1K demonstrate that our approach achieves state-of-the-art reconstruction quality, reaching 0.37 rFID and 26.2 dB PSNR, while maintaining strong semantic alignment to the pretrained VFM. Notably, our Riemannian Flow Matching-based DiT exhibits efficient convergence, achieving a gFID of 3.47 at 80 epochs.

翻译：近期研究探索了使用预训练视觉基础模型（如DINO）构建生成式自编码器，展现出强大的生成性能。然而，现有方法常因高频细节丢失而面临重建保真度受限的问题。本文提出DINO球面自编码器（DINO-SAE），该框架在语义表征与像素级重建之间建立了桥梁。我们的核心洞见在于：对比表征中的语义信息主要编码于特征向量的方向中，而强制严格的幅度匹配会阻碍编码器保留细粒度细节。为此，我们引入层级卷积块嵌入模块以增强局部结构与纹理保持能力，并采用余弦相似度对齐目标函数，在保持语义一致性的同时允许灵活的特征幅度以保留细节。此外，基于自监督学习基础模型表征本质存在于超球面上的观察，我们采用黎曼流匹配方法直接在球面潜在流形上训练扩散Transformer（DiT）。在ImageNet-1K数据集上的实验表明，我们的方法实现了最先进的重建质量，达到0.37 rFID和26.2 dB PSNR，同时保持与预训练视觉基础模型的强语义对齐。值得注意的是，基于黎曼流匹配的DiT展现出高效收敛特性，在80轮训练周期内达到3.47 gFID。