Fashion Outfit Generation via Unified Sequential Composition Models

2026-08-13

AI Summary

This study addresses the challenges of non-monotonic aesthetic compatibility and explosive search spaces in fashion outfit generation by formulating the task as constrained set generation within a finite-horizon Markov Decision Process. We propose the Unified Sequential Combination Model (USCM) and Latent Expansion Monte Carlo Tree Search (LE-MCTS) to jointly model set compatibility and latent intent, effectively balancing local coordination with global structure. This approach represents the first serial decision-making formulation for outfit generation. Extensive experiments on the Polyvore dataset demonstrate state-of-the-art performance, with our method significantly outperforming existing baselines across human preference ratings, automatic aesthetic evaluations, and structural validity metrics. These results validate the efficacy of integrating sequential decision modeling with latent space exploration for generating aesthetically compatible and structurally sound fashion outfits.

Abstract

The task of synthesizing stylistically coherent fashion outfits from massive item libraries, known as fashion outfit generation, remains a non-trivial challenge, primarily due to the non-monotonic and implicit nature of aesthetic compatibility, coupled with the exponentially large combinatorial search space. In this paper, we formalize this task as Constrained Ensemble Generation (CEG) and model it as a finite-horizon deterministic Markov Decision Process. To address CEG in fashion, we propose the Unified Sequential Composition Model (USCM), which jointly models set-level compatibility and latent composition intents. Guided by USCM's learned priors, a Latent Expansion Monte Carlo Tree Search (LE-MCTS) mechanism is proposed to handle item retrieval during composition, balancing local aesthetic synergy with global structural balance. Extensive experiments on the Polyvore Outfits dataset, along with zero-shot evaluations on the iFashion and PolyvoreU datasets, demonstrate that our framework achieves state-of-the-art performance across independent human preference evaluations, automated aesthetic proxies, and structural validity metrics for constrained fashion outfit generation.

Problem

• Fashion Outfit Generation
• Aesthetic Compatibility
• Combinatorial Search Space
• Constrained Ensemble Generation

Innovation

• Unified Sequential Composition Model
• Latent Expansion Monte Carlo Tree Search
• Constrained Ensemble Generation
• Fashion Outfit Generation