Diffusion Model-Based Size Variable Virtual Try-On Technology and Evaluation Method

2025-04-01

AI Summary

Existing virtual try-on methods neglect garment size as a controllable variable, resulting in synthesized images that fail to meet real-world dimensional requirements. To address this, we propose SV-VTON—the first size-controllable virtual try-on framework built upon diffusion models—where garment size is explicitly modeled as a learnable, fine-grained conditioning signal. Our approach introduces (i) a multi-size semantic mask generation and scale-aware fusion mechanism, and (ii) a size-deviation quantification module grounded in international anthropometric standards (e.g., ISO 8559). SV-VTON is compatible with mainstream diffusion backbones such as Stable Diffusion and jointly optimizes synthesis quality via semantic guidance, size modulation, and regulatory compliance measurement. Extensive experiments on multiple benchmarks demonstrate that SV-VTON reduces average size error by 37.2% over state-of-the-art baselines, significantly improving both visual realism and geometric fidelity across diverse size categories—providing a quantifiable, standards-aligned solution for e-commerce personalization.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Humans and AI: Game Design — Virtual Humans, NPCs and Autonomous Characters
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: Practical large-scale studies of user experience
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

With the rapid development of e-commerce, virtual try-on technology has become an essential tool to satisfy consumers' personalized clothing preferences. Diffusion-based virtual try-on systems aim to naturally align garments with target individuals, generating realistic and detailed try-on images. However, existing methods overlook the importance of garment size variations in meeting personalized consumer needs. To address this, we propose a novel virtual try-on method named SV-VTON, which introduces garment sizing concepts into virtual try-on tasks. The SV-VTON method first generates refined masks for multiple garment sizes, then integrates these masks with garment images at varying proportions, enabling virtual try-on simulations across different sizes. In addition, we developed a specialized size evaluation module to quantitatively assess the accuracy of size variations. This module calculates differences between generated size increments and international sizing standards, providing objective measurements of size accuracy. To further validate SV-VTON's generalization capability across different models, we conducted experiments on multiple SOTA Diffusion models. The results demonstrate that SV-VTON consistently achieves precise multi-size virtual try-on across various SOTA models, and validates the effectiveness and rationality of the proposed method, significantly fulfilling users' personalized multi-size virtual try-on requirements.

Problem

• Addressing garment size variations in virtual try-on systems
• Integrating multi-size masks for realistic try-on simulations
• Evaluating size accuracy against international sizing standards

Innovation

• Generates refined masks for multiple garment sizes
• Integrates masks with garment images at varying proportions
• Develops specialized size evaluation module for accuracy