Learning to Measure Change: Fully Convolutional Siamese Metric Networks for Scene Change Detection

TL;DR

This paper introduces CosimNet, a fully convolutional Siamese network with a novel thresholded contrastive loss for robust scene change detection under challenging conditions like illumination and viewpoint variations.

Contribution

The paper proposes a new Siamese metric network with a specialized loss function to improve change detection accuracy amidst noisy and semantic changes.

Findings

Effective in challenging conditions like illumination and viewpoint changes

Robust to noisy changes with the proposed TCL loss

Improves change map quality through integrated metric learning

Abstract

A critical challenge problem of scene change detection is that noisy changes generated by varying illumination, shadows and camera viewpoint make variances of a scene difficult to define and measure since the noisy changes and semantic ones are entangled. Following the intuitive idea of detecting changes by directly comparing dissimilarities between a pair of features, we propose a novel fully Convolutional siamese metric Network(CosimNet) to measure changes by customizing implicit metrics. To learn more discriminative metrics, we utilize contrastive loss to reduce the distance between the unchanged feature pairs and to enlarge the distance between the changed feature pairs. Specifically, to address the issue of large viewpoint differences, we propose Thresholded Contrastive Loss (TCL) with a more tolerant strategy to punish noisy changes. We demonstrate the effectiveness of the proposed approach with experiments on three challenging datasets: CDnet, PCD2015, and VL-CMU-CD. Our approach is robust to lots of challenging conditions, such as illumination changes, large viewpoint difference caused by camera motion and zooming. In addition, we incorporate the distance metric into the segmentation framework and validate the effectiveness through visualization of change maps and feature distribution. The source code is available at https://github.com/gmayday1997/ChangeDet.