DeNISE: Deep Networks for Improved Segmentation Edges

TL;DR

DeNISE is a novel data enhancement method that leverages edge detection and segmentation models to improve boundary accuracy in segmentation masks, demonstrated on aerial building images.

Contribution

DeNISE introduces a versatile, non-end-to-end enhancement technique applicable to various neural networks for better segmentation edge quality.

Findings

Improved building IoU to 78.9% using DeNISE.

Effective in low-resolution, noisy aerial images.

Enhances boundary accuracy without end-to-end training.

Abstract

This paper presents Deep Networks for Improved Segmentation Edges (DeNISE), a novel data enhancement technique using edge detection and segmentation models to improve the boundary quality of segmentation masks. DeNISE utilizes the inherent differences in two sequential deep neural architectures to improve the accuracy of the predicted segmentation edge. DeNISE applies to all types of neural networks and is not trained end-to-end, allowing rapid experiments to discover which models complement each other. We test and apply DeNISE for building segmentation in aerial images. Aerial images are known for difficult conditions as they have a low resolution with optical noise, such as reflections, shadows, and visual obstructions. Overall the paper demonstrates the potential for DeNISE. Using the technique, we improve the baseline results with a building IoU of 78.9%.