Optical coherence tomography (OCT) is a 3D tomographic imaging technique that obtains the internal structure of the specimen. It uses light, and its principle of operation is similar to ultrasound imaging, except that it uses light instead of sound waves and thus achieves much better resolution.
Digital holography is an imaging technique that captures the amplitude and phase of light scattered from a sample. Images can be digitally alatered or refocussed, and aberrations can be corrected. This process is lossless.
Holographic Optical Coherence Tomography combines holographic image processing algorithms with parallelized OCT image acquisition. It is most commonly used in a full-field Fourier domain OCT setup and achieves record-breaking image acquisition rates. In our laboratory, we have an experimental 100 MHz A-line rate holographic OCT setup that acquires phase data without motion artifacts.
The image on the left shows a typical full-field Fourier-domain OCT setup that can be used for retinal imaging.

Optoretinography is one of the most important (future) applications of holographic OCT. Holographic OCT can image the retina and also allows functional contrast of retinal structures, thus enabling phase-based optoretinography, in essance and objective probing of function at the near-cellular or even cellular level. By computationally evaluating the phase in the OCT data, we can obtain changes in the optical path length (size and refractive index) of the photoreceptors and the inner plexiform layer where ganglion and bipolar cells connect to objectively evaluate the function of these layers.
The image on the left shows a retinal image and the morphological and functional signals in various layers that can be extracted from this data.

We are interested in high-performance algorithms and computational methods. For this, we make use of Python and use libraries such as Numpy, Cupy or Pytorch. We also delve into implementing algorithms and software in lower level employing C++, CUDA, or OpenCL. We employ Fourier-transform based algorithms, as well as (non-linear) optimization techniques to achieve our image reconstruction methods.
However, most of our methods are physically motivated. One could say, we are not only doing image procesing but rather data reconstruction solving inverse problems.