Masked Language Prompting for Generative Data Augmentation in Few-shot Fashion Style Recognition
This addresses the problem of limited labeled data for fashion style recognition, offering an incremental improvement in generative data augmentation techniques.
The paper tackles the challenge of constructing datasets for fashion style recognition by proposing Masked Language Prompting (MLP), a generative data augmentation method that masks words in reference captions and uses large language models to create diverse completions, resulting in consistent outperformance over baselines on the FashionStyle14 dataset.
Constructing dataset for fashion style recognition is challenging due to the inherent subjectivity and ambiguity of style concepts. Recent advances in text-to-image models have facilitated generative data augmentation by synthesizing images from labeled data, yet existing methods based solely on class names or reference captions often fail to balance visual diversity and style consistency. In this work, we propose \textbf{Masked Language Prompting (MLP)}, a novel prompting strategy that masks selected words in a reference caption and leverages large language models to generate diverse yet semantically coherent completions. This approach preserves the structural semantics of the original caption while introducing attribute-level variations aligned with the intended style, enabling style-consistent and diverse image generation without fine-tuning. Experimental results on the FashionStyle14 dataset demonstrate that our MLP-based augmentation consistently outperforms class-name and caption-based baselines, validating its effectiveness for fashion style recognition under limited supervision.