This dataset is a benchmark of optical coherence tomography (OCT) volumes of heterogeneous ex vivo animal tissues, released to evaluate implicit neural representations (OCTN) for volumetric reconstruction, sparse acquisition, and 3D tissue representation. The dataset is deliberately diverse in noise level, tissue type, and optical contrast so that it stresses methods—such as INRs—that are susceptible to noise and prone to overfitting easy-to-learn structure.
Volumes were acquired with a low-cost spectral-domain (SD)-OCT system (Lumedica Inc., NC, USA), 1310 nm central wavelength, 110 mm focal length.
