@misc{indiciaee7b4ae1e2c4d,
  title = {Physiognomy Revisited: How Vision-Language Models Encode Sensitive Attributes Through Image Transformation},
  author = {Soumyajit Datta and Arka Dutta and Ashiqur R. KhudaBukhsh},
  year = {2026},
  url = {https://arxiv.org/abs/2610.35910},
  note = {Source identifier: 2610.35910}
}