Xue et al. investigate Fluid-SegNet, a deep learning method designed to better recognize fluid areas in retinal images by targeting differences in the appearance of these regions. Xue et al tested the Fluid-SegNet deep learning model on three public Optical Coherence Tomography (OCT) datasets– UMN, AROI, & OIMHS– to evaluate its accuracy in detecting fluid regions in retinal images. Xue et al. found that Fluid-SegNet accurately recognized fluid regions in retinal OCT images as it showed high dice scores across all datasets. High Dice scores indicated that the model’s predicted fluid regions matched the actual ones. Overall, these results showed that Fluid-SegNet works better than other models at finding small and hard to see fluid areas in eye scans. Xue et al. also found that adding certain features to the model helped it more accurately detect the fluid regions. However, the model sometimes mistook nearby fluid areas as one big area and had trouble with blurry images. Therefore future research on the model should look into fixing this issue.