DECOR:Decomposition and Projection of Text Embeddings for Text-to-Image Customization
- URL: http://arxiv.org/abs/2412.09169v1
- Date: Thu, 12 Dec 2024 10:59:44 GMT
- Title: DECOR:Decomposition and Projection of Text Embeddings for Text-to-Image Customization
- Authors: Geonhui Jang, Jin-Hwa Kim, Yong-Hyun Park, Junho Kim, Gayoung Lee, Yonghyun Jeong,
- Abstract summary: This study decomposes the text embedding matrix and conducts a component analysis to understand the embedding space geometry.
We propose DECOR, which projects text embeddings onto a vector space to undesired token vectors.
Experimental results demonstrate that DECOR outperforms state-of-the-art customization models.
- Score: 15.920735314050296
- License:
- Abstract: Text-to-image (T2I) models can effectively capture the content or style of reference images to perform high-quality customization. A representative technique for this is fine-tuning using low-rank adaptations (LoRA), which enables efficient model customization with reference images. However, fine-tuning with a limited number of reference images often leads to overfitting, resulting in issues such as prompt misalignment or content leakage. These issues prevent the model from accurately following the input prompt or generating undesired objects during inference. To address this problem, we examine the text embeddings that guide the diffusion model during inference. This study decomposes the text embedding matrix and conducts a component analysis to understand the embedding space geometry and identify the cause of overfitting. Based on this, we propose DECOR, which projects text embeddings onto a vector space orthogonal to undesired token vectors, thereby reducing the influence of unwanted semantics in the text embeddings. Experimental results demonstrate that DECOR outperforms state-of-the-art customization models and achieves Pareto frontier performance across text and visual alignment evaluation metrics. Furthermore, it generates images more faithful to the input prompts, showcasing its effectiveness in addressing overfitting and enhancing text-to-image customization.
Related papers
- Bridging the Gap: Aligning Text-to-Image Diffusion Models with Specific Feedback [5.415802995586328]
Learning from feedback has been shown to enhance the alignment between text prompts and images in text-to-image diffusion models.
We propose an efficient fine-turning method with specific reward objectives, including three stages.
Experimental results on this benchmark show that our model outperforms other SOTA methods in both alignment and fidelity.
arXiv Detail & Related papers (2024-11-28T09:56:28Z) - Object-Attribute Binding in Text-to-Image Generation: Evaluation and Control [58.37323932401379]
Current diffusion models create images given a text prompt as input but struggle to correctly bind attributes mentioned in the text to the right objects in the image.
We propose focused cross-attention (FCA) that controls the visual attention maps by syntactic constraints found in the input sentence.
We show substantial improvements in T2I generation and especially its attribute-object binding on several datasets.
arXiv Detail & Related papers (2024-04-21T20:26:46Z) - Contrastive Prompts Improve Disentanglement in Text-to-Image Diffusion
Models [68.47333676663312]
We show a simple modification of classifier-free guidance can help disentangle image factors in text-to-image models.
The key idea of our method, Contrastive Guidance, is to characterize an intended factor with two prompts that differ in minimal tokens.
We illustrate whose benefits in three scenarios: (1) to guide domain-specific diffusion models trained on an object class, (2) to gain continuous, rig-like controls for text-to-image generation, and (3) to improve the performance of zero-shot image editors.
arXiv Detail & Related papers (2024-02-21T03:01:17Z) - UDiffText: A Unified Framework for High-quality Text Synthesis in
Arbitrary Images via Character-aware Diffusion Models [25.219960711604728]
This paper proposes a novel approach for text image generation, utilizing a pre-trained diffusion model.
Our approach involves the design and training of a light-weight character-level text encoder, which replaces the original CLIP encoder.
By employing an inference stage refinement process, we achieve a notably high sequence accuracy when synthesizing text in arbitrarily given images.
arXiv Detail & Related papers (2023-12-08T07:47:46Z) - Improving Compositional Text-to-image Generation with Large
Vision-Language Models [26.202725136839632]
compositional text-to-image models frequently encounter difficulties in generating high-quality images that align with input texts.
We employ large vision-language models (LVLMs) for multi-dimensional assessment of the alignment between generated images and their corresponding input texts.
Our experimental results validate that the proposed methodology significantly improves text-image alignment in compositional image generation.
arXiv Detail & Related papers (2023-10-10T05:09:05Z) - Dynamic Prompt Learning: Addressing Cross-Attention Leakage for
Text-Based Image Editing [23.00202969969574]
We propose Dynamic Prompt Learning (DPL) to force cross-attention maps to focus on correct noun words in the text prompt.
We show improved prompt editing results for Word-Swap, Prompt Refinement, and Attention Re-weighting, especially for complex multi-object scenes.
arXiv Detail & Related papers (2023-09-27T13:55:57Z) - Zero-shot spatial layout conditioning for text-to-image diffusion models [52.24744018240424]
Large-scale text-to-image diffusion models have significantly improved the state of the art in generative image modelling.
We consider image generation from text associated with segments on the image canvas, which combines an intuitive natural language interface with precise spatial control over the generated content.
We propose ZestGuide, a zero-shot segmentation guidance approach that can be plugged into pre-trained text-to-image diffusion models.
arXiv Detail & Related papers (2023-06-23T19:24:48Z) - iEdit: Localised Text-guided Image Editing with Weak Supervision [53.082196061014734]
We propose a novel learning method for text-guided image editing.
It generates images conditioned on a source image and a textual edit prompt.
It shows favourable results against its counterparts in terms of image fidelity, CLIP alignment score and qualitatively for editing both generated and real images.
arXiv Detail & Related papers (2023-05-10T07:39:14Z) - Harnessing the Spatial-Temporal Attention of Diffusion Models for
High-Fidelity Text-to-Image Synthesis [59.10787643285506]
Diffusion-based models have achieved state-of-the-art performance on text-to-image synthesis tasks.
One critical limitation of these models is the low fidelity of generated images with respect to the text description.
We propose a new text-to-image algorithm that adds explicit control over spatial-temporal cross-attention in diffusion models.
arXiv Detail & Related papers (2023-04-07T23:49:34Z) - High-Fidelity Guided Image Synthesis with Latent Diffusion Models [50.39294302741698]
The proposed approach outperforms the previous state-of-the-art by over 85.32% on the overall user satisfaction scores.
Human user study results show that the proposed approach outperforms the previous state-of-the-art by over 85.32% on the overall user satisfaction scores.
arXiv Detail & Related papers (2022-11-30T15:43:20Z) - CLIP-PAE: Projection-Augmentation Embedding to Extract Relevant Features for a Disentangled, Interpretable, and Controllable Text-Guided Face Manipulation [4.078926358349661]
Contrastive Language-Image Pre-Training (CLIP) bridges images and text by embedding them into a joint latent space.
Due to the discrepancy between image and text embeddings in the joint space, using text embeddings as the optimization target often introduces undesired artifacts in the resulting images.
We introduce CLIP Projection-Augmentation Embedding (PAE) as an optimization target to improve the performance of text-guided image manipulation.
arXiv Detail & Related papers (2022-10-08T05:12:25Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.