Saved in:
Bibliographic Details
Main Authors: Esteves, Carlos, Suhail, Mohammed, Makadia, Ameesh
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.09607
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915335781744640
author Esteves, Carlos
Suhail, Mohammed
Makadia, Ameesh
author_facet Esteves, Carlos
Suhail, Mohammed
Makadia, Ameesh
contents Image tokenizers map images to sequences of discrete tokens, and are a crucial component of autoregressive transformer-based image generation. The tokens are typically associated with spatial locations in the input image, arranged in raster scan order, which is not ideal for autoregressive modeling. In this paper, we propose to tokenize the image spectrum instead, obtained from a discrete wavelet transform (DWT), such that the sequence of tokens represents the image in a coarse-to-fine fashion. Our tokenizer brings several advantages: 1) it leverages that natural images are more compressible at high frequencies, 2) it can take and reconstruct images of different resolutions without retraining, 3) it improves the conditioning for next-token prediction -- instead of conditioning on a partial line-by-line reconstruction of the image, it takes a coarse reconstruction of the full image, 4) it enables partial decoding where the first few generated tokens can reconstruct a coarse version of the image, 5) it enables autoregressive models to be used for image upsampling. We evaluate the tokenizer reconstruction metrics as well as multiscale image generation, text-guided image upsampling and editing.
format Preprint
id arxiv_https___arxiv_org_abs_2412_09607
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Spectral Image Tokenizer
Esteves, Carlos
Suhail, Mohammed
Makadia, Ameesh
Computer Vision and Pattern Recognition
Machine Learning
Image tokenizers map images to sequences of discrete tokens, and are a crucial component of autoregressive transformer-based image generation. The tokens are typically associated with spatial locations in the input image, arranged in raster scan order, which is not ideal for autoregressive modeling. In this paper, we propose to tokenize the image spectrum instead, obtained from a discrete wavelet transform (DWT), such that the sequence of tokens represents the image in a coarse-to-fine fashion. Our tokenizer brings several advantages: 1) it leverages that natural images are more compressible at high frequencies, 2) it can take and reconstruct images of different resolutions without retraining, 3) it improves the conditioning for next-token prediction -- instead of conditioning on a partial line-by-line reconstruction of the image, it takes a coarse reconstruction of the full image, 4) it enables partial decoding where the first few generated tokens can reconstruct a coarse version of the image, 5) it enables autoregressive models to be used for image upsampling. We evaluate the tokenizer reconstruction metrics as well as multiscale image generation, text-guided image upsampling and editing.
title Spectral Image Tokenizer
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2412.09607