Text-to-Person Retrieval With Disentangled Identity and Apparel Representations in Multimodal Vision Systems. JAAIR 2026, 1 (1).