{
  "annotation_id": "03c88c81-ecf5-4608-90ee-435bb926b6e4",
  "date_created": "2026-03-02T18:01:35.043000Z",
  "date_modified": "2026-03-02T18:01:35.043000Z",
  "file_hash": "7a5634a502b4688ced7b950000bc0f170297903f21228fabe9805038d9959aa9",
  "private": false,
  "record": {
    "abstract": "Psychophysical studies suggest that face recognition takes place in a narrow\nband of low spatial frequencies (``critical band\u0027\u0027). Here, we examined the\nrecognition performance of an artificial face recognition system as a function\nof the size of the input images. Recognition performance was quantified with\nthree discriminability measures: Fisher Linear Discriminant Analysis, non\nParametric Discriminant Analysis, and mutual information. All of the three\nmeasures revealed a maximum at the same image sizes. Since spatial frequency\ncontent is a function of image size, our data consistently predict the range of\npsychophysical found frequencies. Our results therefore support the notion that\nthe critical band of spatial frequencies for face recognition in humans and\nmachines follows from inherent properties of face images.",
    "arxiv_id": "q-bio/0612001",
    "authors": [
      "Matthias S. Keil",
      "Agata Lapedriza",
      "David Masip",
      "Jordi Vitria"
    ],
    "categories": [
      "q-bio.NC"
    ],
    "title": "Face Recognition in the Machine Reveals Properties of Human Face   Recognition",
    "url": "https://arxiv.org/abs/q-bio/0612001"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "7a2b87e1-208d-4524-8d60-8cee9c21389f",
    "id": "arXiv Dataset IDs",
    "type": "Model",
    "variant": "snapshot-2026-03-01",
    "version": "0.1.0"
  },
  "user_id": 1000002
}