POLARIS: Projection-Orthogonal Least Squares for Robust and Adaptive Inversion in Diffusion Models

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Wenshuo, Li, Haosen, Liang, Shaofeng, Wang, Lei, Jia, Haozhe, Yuan, Kaishen, Wu, Jieming, Tian, Bowen, Yue, Yutao
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914174661033984
author Chen, Wenshuo
Li, Haosen
Liang, Shaofeng
Wang, Lei
Jia, Haozhe
Yuan, Kaishen
Wu, Jieming
Tian, Bowen
Yue, Yutao
author_facet Chen, Wenshuo
Li, Haosen
Liang, Shaofeng
Wang, Lei
Jia, Haozhe
Yuan, Kaishen
Wu, Jieming
Tian, Bowen
Yue, Yutao
contents The Inversion-Denoising Paradigm, which is based on diffusion models, excels in diverse image editing and restoration tasks. We revisit its mechanism and reveal a critical, overlooked factor in reconstruction degradation: the approximate noise error. This error stems from approximating the noise at step t with the prediction at step t-1, resulting in severe error accumulation throughout the inversion process. We introduce Projection-Orthogonal Least Squares for Robust and Adaptive Inversion (POLARIS), which reformulates inversion from an error-compensation problem into an error-origin problem. Rather than optimizing embeddings or latent codes to offset accumulated drift, POLARIS treats the guidance scale ω as a step-wise variable and derives a mathematically grounded formula to minimize inversion error at each step. Remarkably, POLARIS improves inversion latent quality with just one line of code. With negligible performance overhead, it substantially mitigates noise approximation errors and consistently improves the accuracy of downstream tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2512_00369
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle POLARIS: Projection-Orthogonal Least Squares for Robust and Adaptive Inversion in Diffusion Models
Chen, Wenshuo
Li, Haosen
Liang, Shaofeng
Wang, Lei
Jia, Haozhe
Yuan, Kaishen
Wu, Jieming
Tian, Bowen
Yue, Yutao
Computer Vision and Pattern Recognition
The Inversion-Denoising Paradigm, which is based on diffusion models, excels in diverse image editing and restoration tasks. We revisit its mechanism and reveal a critical, overlooked factor in reconstruction degradation: the approximate noise error. This error stems from approximating the noise at step t with the prediction at step t-1, resulting in severe error accumulation throughout the inversion process. We introduce Projection-Orthogonal Least Squares for Robust and Adaptive Inversion (POLARIS), which reformulates inversion from an error-compensation problem into an error-origin problem. Rather than optimizing embeddings or latent codes to offset accumulated drift, POLARIS treats the guidance scale ω as a step-wise variable and derives a mathematically grounded formula to minimize inversion error at each step. Remarkably, POLARIS improves inversion latent quality with just one line of code. With negligible performance overhead, it substantially mitigates noise approximation errors and consistently improves the accuracy of downstream tasks.
title POLARIS: Projection-Orthogonal Least Squares for Robust and Adaptive Inversion in Diffusion Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.00369