Aria App | LittleOne Edition, an All-in-One Application for Processing Medical Images and Developing U-Net Models

The efficient segmentation of regions of interest is a crucial step for various medical applications, as it is widely used for monitoring disease progression and evaluating various conditions. However, the traditional manual segmentation by trained experts is a time consuming and repetitive process. Various approaches have been proposed in recent years such as and not limited to the use of neural networks and deep learning. Inspired by the achievements of artificial intelligence in demanding computer vision tasks we present the Aria App | LittleOne edition, featuring a GUI where image processing, network development and evaluation can be conducted. In this paper, we attempt to address the challenge of identifying retinal vessels in fundus images which is accomplished using the U-Net network. Experimental results conducted within the app on the DRIVE dataset demonstrate comparable performance to recent research with a 94.9% accuracy, 93.2% sensitivity, 95.1% specificity and 94.20% AUC in the total.