| { |
| "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." |
| } |
| ] |
| } |
|
|
