NumeriKontrol: Adding Numeric Control to Diffusion Transformers for Instruction-based Image Editing

2025-11-28

AI Summary

Existing instruction-based image editing methods lack fine-grained, numerically precise control over attribute modification strength. To address this, we propose Numeric Adapter—a plug-and-play numerical encoding module that embeds continuous scalar strength parameters into diffusion Transformer architectures, enabling joint text-instruction and numeric-strength conditioning. Leveraging a task-decoupled design and the high-fidelity, real-scale supervised synthetic dataset CAT, our method achieves, for the first time, multi-attribute disentanglement, zero-shot generalization, arbitrary-order input handling, and stable, continuous strength modulation. Training data integrate photorealistic renderings with DSLR-captured images. Experiments demonstrate accurate, consistent, and controllable scaling across diverse attributes—including color, shape, and size—significantly improving both precision in strength specification and output stability.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Adversarial Learning & Robustness

Application Category

• User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertising
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Instruction-based image editing enables intuitive manipulation through natural language commands. However, text instructions alone often lack the precision required for fine-grained control over edit intensity. We introduce NumeriKontrol, a framework that allows users to precisely adjust image attributes using continuous scalar values with common units. NumeriKontrol encodes numeric editing scales via an effective Numeric Adapter and injects them into diffusion models in a plug-and-play manner. Thanks to a task-separated design, our approach supports zero-shot multi-condition editing, allowing users to specify multiple instructions in any order. To provide high-quality supervision, we synthesize precise training data from reliable sources, including high-fidelity rendering engines and DSLR cameras. Our Common Attribute Transform (CAT) dataset covers diverse attribute manipulations with accurate ground-truth scales, enabling NumeriKontrol to function as a simple yet powerful interactive editing studio. Extensive experiments show that NumeriKontrol delivers accurate, continuous, and stable scale control across a wide range of attribute editing scenarios. These contributions advance instruction-based image editing by enabling precise, scalable, and user-controllable image manipulation.

Problem

• Enables precise numeric control of image edit intensity
• Supports zero-shot multi-condition editing through natural language
• Provides continuous scale manipulation across diverse attribute scenarios

Innovation

• Numeric Adapter encodes continuous scalar values
• Plug-and-play injection into diffusion models
• Task-separated design enables zero-shot multi-condition editing