Shapefit: Turn irregular shape masks into clean polygons (github.com)

Shapefit, a new neural network tool, has been introduced to refine irregular shape masks into clean polygons, targeting issues like jagged edges and gaps. By accepting input masks with various imperfections, Shapefit generates precise polygonal representations, enhancing applications in architectural layouts and spatial analysis. The open-source project can be easily implemented with a few Python commands, requiring libraries like numpy, PyTorch, and OpenCV, which streamline the installation and testing processes. This development is significant for the AI/ML community as it provides a more effective alternative to traditional polygon approximation algorithms, such as the Douglas-Peucker method. Performance metrics indicate that Shapefit consistently surpasses its predecessor in terms of Intersection over Union (IoU), corner accuracy, and vertex F1 scores across diverse shape types. For instance, in tests with 500 various shapes, Shapefit achieved an IoU of 0.960 compared to Douglas-Peucker's highest score of 0.910, showcasing its potential to improve shape simplification tasks in machine learning and computer vision applications.