DeepImageSegmentationApp: Deep Learning Application for Image Segmentation

Abstract

There are many methods to examine a specific region or object in images. One of the most important of these methods is image segmentation. Image segmentation involves dividing images (or video frames) into multiple sections or objects. There are many different model architectures developed in the field of image segmentation. In this study, a deep learning-based image segmentation application interface has been developed. The performance of the proposed application has been analyzed on the Covid19 dataset obtained from Kaggle. The performance results of the application are presented comparatively on the U-NET and V-NET models with known accuracy for different system parameters. In the analysis results, it is clearly seen that the V-NET architecture is better than the U-NET architecture. The developed application environment has revealed the difference between the models and the usability of the application environment. This standalone software can be downloaded at: https://github.com/lbayrak/DeepImageSegmentationApp.