Training-Free Disentangled Text-Guided Image Editing via Sparse Latent Constraints

TL;DR

This paper proposes a training-free, disentangled image editing method guided by text, using sparse latent constraints to improve control and reduce attribute entanglement in face editing tasks.

Contribution

It introduces a sparsity-based regularization to the PPE framework, enhancing disentanglement and control in text-guided image editing without additional training.

Findings

Reduces unintended attribute changes in face editing.

Enforces more focused and controlled image edits.

Preserves facial identity during attribute modifications.

Abstract

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.