@misc{indiciae55ac0e45409b,
  title = {CaptionFuse: Training-Free Composed Product Retrieval Via VLM Intent Captioning and Weighted Embedding Fusion},
  author = {Zhao Gao and Mina Metias and Mohamed Noordeen Alaudeen and Bassel Saleh},
  year = {2026},
  url = {https://arxiv.org/abs/2610.26661},
  note = {Source identifier: 2610.26661}
}