Saved in:
Bibliographic Details
Main Authors: Shabrina, Mutiara, Putri, Nova Kurnia, Ferdiansyah, Jefri Satria, Dewi, Sabita Khansa, Yudistira, Novanto
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.21637
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912789213216768
author Shabrina, Mutiara
Putri, Nova Kurnia
Ferdiansyah, Jefri Satria
Dewi, Sabita Khansa
Yudistira, Novanto
author_facet Shabrina, Mutiara
Putri, Nova Kurnia
Ferdiansyah, Jefri Satria
Dewi, Sabita Khansa
Yudistira, Novanto
contents Text-driven image manipulation often suffers from attribute entanglement, where modifying a target attribute (e.g., adding bangs) unintentionally alters other semantic properties such as identity or appearance. The Predict, Prevent, and Evaluate (PPE) framework addresses this issue by leveraging pre-trained vision-language models for disentangled editing. In this work, we analyze the PPE framework, focusing on its architectural components, including BERT-based attribute prediction and StyleGAN2-based image generation on the CelebA-HQ dataset. Through empirical analysis, we identify a limitation in the original regularization strategy, where latent updates remain dense and prone to semantic leakage. To mitigate this issue, we introduce a sparsity-based constraint using L1 regularization on latent space manipulation. Experimental results demonstrate that the proposed approach enforces more focused and controlled edits, effectively reducing unintended changes in non-target attributes while preserving facial identity.
format Preprint
id arxiv_https___arxiv_org_abs_2512_21637
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Training-Free Disentangled Text-Guided Image Editing via Sparse Latent Constraints
Shabrina, Mutiara
Putri, Nova Kurnia
Ferdiansyah, Jefri Satria
Dewi, Sabita Khansa
Yudistira, Novanto
Computer Vision and Pattern Recognition
Text-driven image manipulation often suffers from attribute entanglement, where modifying a target attribute (e.g., adding bangs) unintentionally alters other semantic properties such as identity or appearance. The Predict, Prevent, and Evaluate (PPE) framework addresses this issue by leveraging pre-trained vision-language models for disentangled editing. In this work, we analyze the PPE framework, focusing on its architectural components, including BERT-based attribute prediction and StyleGAN2-based image generation on the CelebA-HQ dataset. Through empirical analysis, we identify a limitation in the original regularization strategy, where latent updates remain dense and prone to semantic leakage. To mitigate this issue, we introduce a sparsity-based constraint using L1 regularization on latent space manipulation. Experimental results demonstrate that the proposed approach enforces more focused and controlled edits, effectively reducing unintended changes in non-target attributes while preserving facial identity.
title Training-Free Disentangled Text-Guided Image Editing via Sparse Latent Constraints
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.21637