Optical coherence tomography (OCT) is an imaging device widely used in ophthalmology, but attempts are underway to obtain images with contrast different from conventional OCT images. The pixel value of conventional OCT images corresponds the intensity of scattered light, which is not very suitable for evaluating tissue characteristics. Therefore, methods to generate contrast focusing on tissue density, anisotropic scatterers, and strong scatterers have been investigated by computational processing such as numerical simulation based on physical models and deep learning. These new contrast generation methods are expected to visualize information on tissue characteristics and diseases that cannot be obtained with conventional OCT images, and to be used as new imaging biomarkers. In particular, the development of contrast generators using deep learning has the potential to solve complex inverse problems and generate contrasts that have been difficult to achieve. The development of these new contrast generators and imaging biomarkers is expected to further expand the applications of OCT imaging.