Comprehensive Topcon 3D-OCT Dataset: Inclusive of Normal and Abnormal OCT Volumes

This dataset consists of 948 OCT volumes, with 413 categorized as Normal and 535 as Abnormal. These volumes were acquired using the Topcon 3D OCT-1000 imaging system at the Ophthalmology Department of Feiz Hospital, Isfahan, Iran.

Each volume in the dataset contains a variable number of B-scan images. The resolution of each B-scan is 650 x 512 pixels. In total, there are 57,171 B-scans. Among these, 33,314 B-scans are categorized as Normal, while 23,857 are categorized as Abnormal.

P. G. Daneshmand and H. Rabbani, “Total Variation Regularized Tensor Ring Decomposition for OCT Image Denoising and Super-resolution,” (submitted to Computers in biology and medicine), 2024.

P. G. Daneshmand and H. Rabbani, “Tensor Ring Decomposition Guided Dictionary Learning for OCT Image Denoising,” IEEE Transactions on Medical Imaging, 2024.

Download Full Database