{
  "field_id": "F040",
  "field": "computational imaging",
  "directions": [
    {
      "direction_id": "G0",
      "name": "Design of imaging measurements",
      "scope": "Focus on how measurements are acquired: design optical encodings, illumination, and sampling schemes, or use feedback during acquisition to select subsequent measurements, to improve image recovery or downstream performance."
    },
    {
      "direction_id": "G1",
      "name": "Measurement-conditioned inverse inference",
      "scope": "Study how to combine imaging measurements with reusable priors for measurement-consistent reconstruction or posterior inference; fully encompass denoiser-based iterative priors, inference of probability distributions and uncertainty, and measurement-guided generative reconstruction."
    },
    {
      "direction_id": "G2",
      "name": "Optical wavefront recovery",
      "scope": "Recover optical phase or complex-valued structure that was not directly measured from diffraction, intensity, or holographic observations."
    },
    {
      "direction_id": "G3",
      "name": "Reconstruction from compressive imaging measurements",
      "scope": "Recover spatial images, spectral images, or video frame sequences from compressed measurements under a given coded or multiplexed acquisition."
    },
    {
      "direction_id": "G4",
      "name": "Recovery of unsampled image detail",
      "scope": "When color, spectral, or polarization values, or high-resolution spatial detail, have not been directly sampled at target locations, infer the needed content from existing image observations."
    },
    {
      "direction_id": "G5",
      "name": "MRI and tomographic measurement inversion",
      "scope": "Recover images, image sequences, or internal volumes from incomplete k-space, projection, or related tomographic measurements, while addressing the corresponding acquisition consistency, geometry, and forward-model issues."
    },
    {
      "direction_id": "G6",
      "name": "Restoration of degraded recorded images",
      "scope": "With existing images as the primary objects of restoration, recover content and clarity degraded by noise, blur, optical aberrations, or artifacts following initial tomographic reconstruction."
    },
    {
      "direction_id": "G7",
      "name": "Camera radiometry and image processing",
      "scope": "Obtain usable camera-image appearance from exposure or sensor-domain observations, encompassing multi-exposure HDR reconstruction, RAW and photo-domain processing, and enhancement of visibility and color in existing low-light images."
    },
    {
      "direction_id": "G8",
      "name": "Visibility recovery through atmospheric haze",
      "scope": "Recover scene visibility and appearance from images affected by atmospheric haze, including spatially varying and nighttime conditions."
    },
    {
      "direction_id": "G9",
      "name": "Recovery of scene structure and appearance",
      "scope": "Infer scene depth, geometry, appearance, or renderable representations from direct or indirect imaging evidence; fully encompass radiance-field novel-view reconstruction, decomposition into physical shape, materials, and illumination, depth recovery, and reconstruction of hidden scenes outside the direct line of sight."
    },
    {
      "direction_id": "G10",
      "name": "Hyperspectral material-component inference",
      "scope": "Infer constituent material endmember spectra and their abundances from observed mixed hyperspectral signals, rather than aiming to recover missing spectral intensities."
    }
  ]
}