Multiscale Softmax Cross Entropy for Fovea Localization on Color Fundus Photography

TL;DR

This paper introduces a multiscale softmax cross entropy method for fovea localization in color fundus images, improving coordinate prediction accuracy by modifying the loss and activation functions.

Contribution

The paper proposes a novel multiscale softmax cross entropy approach for coordinate regression in ophthalmic image analysis, outperforming traditional methods.

Findings

Multiscale softmax cross entropy improves localization accuracy.

The method outperforms vanilla softmax and MSE with sigmoid.

Enhanced coordinate predictions in fundus photography.

Abstract

Fovea localization is one of the most popular tasks in ophthalmic medical image analysis, where the coordinates of the center point of the macula lutea, i.e. fovea centralis, should be calculated based on color fundus images. In this work, we treat the localization problem as a classification task, where the coordinates of the x- and y-axis are considered as the target classes. Moreover, the combination of the softmax activation function and the cross entropy loss function is modified to its multiscale variation to encourage the predicted coordinates to be located closely to the ground-truths. Based on color fundus photography images, we empirically show that the proposed multiscale softmax cross entropy yields better performance than the vanilla version and than the mean squared error loss with sigmoid activation, which provides a novel approach for coordinate regression.