Preserve or Modify? Context-Aware Evaluation for Balancing Preservation and Modification in Text-Guided Image Editing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Yoonjeon, Ryu, Soohyun, Jung, Yeonsung, Lee, Hyunkoo, Kim, Joowon, Yang, June Yong, Hwang, Jaeryong, Yang, Eunho
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917962539073536
author Kim, Yoonjeon
Ryu, Soohyun
Jung, Yeonsung
Lee, Hyunkoo
Kim, Joowon
Yang, June Yong
Hwang, Jaeryong
Yang, Eunho
author_facet Kim, Yoonjeon
Ryu, Soohyun
Jung, Yeonsung
Lee, Hyunkoo
Kim, Joowon
Yang, June Yong
Hwang, Jaeryong
Yang, Eunho
contents The development of vision-language and generative models has significantly advanced text-guided image editing, which seeks the preservation of core elements in the source image while implementing modifications based on the target text. However, existing metrics have a context-blindness problem, indiscriminately applying the same evaluation criteria on completely different pairs of source image and target text, biasing towards either modification or preservation. Directional CLIP similarity, the only metric that considers both source image and target text, is also biased towards modification aspects and attends to irrelevant editing regions of the image. We propose AugCLIP, a context-aware metric that adaptively coordinates preservation and modification aspects, depending on the specific context of a given source image and target text. This is done by deriving the CLIP representation of an ideally edited image, that preserves the source image with necessary modifications to align with target text. More specifically, using a multi-modal large language model, AugCLIP augments the textual descriptions of the source and target, then calculates a modification vector through a hyperplane that separates source and target attributes in CLIP space. Extensive experiments on five benchmark datasets, encompassing a diverse range of editing scenarios, show that AugCLIP aligns remarkably well with human evaluation standards, outperforming existing metrics. The code is available at https://github.com/augclip/augclip_eval.
format Preprint
id arxiv_https___arxiv_org_abs_2410_11374
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Preserve or Modify? Context-Aware Evaluation for Balancing Preservation and Modification in Text-Guided Image Editing
Kim, Yoonjeon
Ryu, Soohyun
Jung, Yeonsung
Lee, Hyunkoo
Kim, Joowon
Yang, June Yong
Hwang, Jaeryong
Yang, Eunho
Computer Vision and Pattern Recognition
Artificial Intelligence
The development of vision-language and generative models has significantly advanced text-guided image editing, which seeks the preservation of core elements in the source image while implementing modifications based on the target text. However, existing metrics have a context-blindness problem, indiscriminately applying the same evaluation criteria on completely different pairs of source image and target text, biasing towards either modification or preservation. Directional CLIP similarity, the only metric that considers both source image and target text, is also biased towards modification aspects and attends to irrelevant editing regions of the image. We propose AugCLIP, a context-aware metric that adaptively coordinates preservation and modification aspects, depending on the specific context of a given source image and target text. This is done by deriving the CLIP representation of an ideally edited image, that preserves the source image with necessary modifications to align with target text. More specifically, using a multi-modal large language model, AugCLIP augments the textual descriptions of the source and target, then calculates a modification vector through a hyperplane that separates source and target attributes in CLIP space. Extensive experiments on five benchmark datasets, encompassing a diverse range of editing scenarios, show that AugCLIP aligns remarkably well with human evaluation standards, outperforming existing metrics. The code is available at https://github.com/augclip/augclip_eval.
title Preserve or Modify? Context-Aware Evaluation for Balancing Preservation and Modification in Text-Guided Image Editing
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2410.11374