CVNov 25, 2024

Factorized Visual Tokenization and Generation

arXiv:2411.16681v210 citationsh-index: 9
Originality Highly original
AI Analysis

This addresses a critical bottleneck in image generation for AI researchers, offering a scalable solution to improve tokenizer performance, though it is incremental on existing VQ methods.

The paper tackles the scalability challenge of VQ-based visual tokenizers by introducing Factorized Quantization (FQ), which decomposes large codebooks into sub-codebooks with disentanglement regularization and representation learning, achieving state-of-the-art reconstruction quality and enabling effective auto-regressive image generation.

Visual tokenizers are fundamental to image generation. They convert visual data into discrete tokens, enabling transformer-based models to excel at image generation. Despite their success, VQ-based tokenizers like VQGAN face significant limitations due to constrained vocabulary sizes. Simply expanding the codebook often leads to training instability and diminishing performance gains, making scalability a critical challenge. In this work, we introduce Factorized Quantization (FQ), a novel approach that revitalizes VQ-based tokenizers by decomposing a large codebook into multiple independent sub-codebooks. This factorization reduces the lookup complexity of large codebooks, enabling more efficient and scalable visual tokenization. To ensure each sub-codebook captures distinct and complementary information, we propose a disentanglement regularization that explicitly reduces redundancy, promoting diversity across the sub-codebooks. Furthermore, we integrate representation learning into the training process, leveraging pretrained vision models like CLIP and DINO to infuse semantic richness into the learned representations. This design ensures our tokenizer captures diverse semantic levels, leading to more expressive and disentangled representations. Experiments show that the proposed FQGAN model substantially improves the reconstruction quality of visual tokenizers, achieving state-of-the-art performance. We further demonstrate that this tokenizer can be effectively adapted into auto-regressive image generation. https://showlab.github.io/FQGAN

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes