@article{fdi:010086656,
  title = {{H}olographic reconstruction enhancement via unpaired image-to-image translation},
  author = {{S}cherrer, {R}. and {S}elmaoui-folcher, {N}. and {Q}uiniou, {T}. and {J}auffrais, {T}. and {L}emonnier, {H}. and {B}onnet, {S}ophie},
  editor = {},
  language = {{ENG}},
  abstract = {{D}igital holographic microscopy is an imaging process that encodes the 3{D} information of a sample into a single 2{D} hologram. {T}he holographic reconstruction that decodes the hologram is conventionally based on the diffraction formula and involves various iterative steps in order to recover the lost phase information of the hologram. {I}n the past few years, the deep-learning-based model has shown great potential to perform holographic reconstruction directly on a single hologram. {H}owever, preparing a large and high-quality dataset to train the models remains a challenge, especially when the holographic reconstruction images that serve as ground truth are difficult to obtain and can have a deteriorated quality due to various interferences of the imaging device. {A} cycle generative adversarial network is first trained with unpaired brightfield microscope images to restore the visual quality of the holographic reconstructions. {T}he enhanced holographic reconstructions then serve as ground truth for the supervised learning of a {U}-{N}et that performs the holographic reconstruction on a single hologram. {T}he proposed method was evalu-ated on plankton images and could also be applied to achieve super-resolution or colorization of the holographic reconstructions.},
  keywords = {},
  booktitle = {},
  journal = {{A}pplied {O}ptics},
  volume = {61},
  numero = {33},
  pages = {9807--9816},
  ISSN = {1559-128{X}},
  year = {2022},
  DOI = {10.1364/ao.471131},
  URL = {https://www.documentation.ird.fr/hor/fdi:010086656},
}