Suitable Embedding to Find Similarity Between Left and Right Retinas of a Person

It is often argued among biometric researchers that the left and right retinas of the same person are as different as the retinas of two different persons. However, in our previous work we showed that by looking at retinal images human volunteers can find some similarity and tell whether a pair of the left and right retinal images belongs to a single person or to two different persons. We also showed that using cosine similarity between the embedded vectors of two retinal images it is possible to find the correct right retinal image for a left retinal image more than what could be expected by chance. In this paper we investigate which type of embedding can give us better cosine similarity measurement between the left and right retinal images of a person.