Do we even need clinicians be able to distinguish male and female retinas?
The task is meaningless. Yes, there might be some interesting facts in discovery how male and female retinas are different. It could even lead to differentiated treatments. But ML hasn't provided any clues regarding this and therefore it is not that deep.
It's not about differentiating male and female retinas (that's just a PoC to publish the paper), it's about using ML to find something useful in data (e.g. signs of a disease) which might be hard to see for humans otherwise.
>Clinicians are currently unaware of distinct retinal feature variations between males and females, highlighting the importance of model explainability for this task.
