Guided Anisotropic Diffusion and Iterative Learning for Weakly Supervised Change Detection
This work addresses weakly supervised learning for change detection, offering incremental improvements in handling noisy datasets from open sources.
The paper tackles the problem of noisy and unreliable training data in large-scale change detection by proposing an iterative learning method that extracts useful information from open vector data to train a fully convolutional network, achieving performance surpassing naive supervised learning. It also introduces a guided anisotropic diffusion algorithm that improves semantic segmentation through edge-preserving filtering, used alongside iterative training to enhance results.
Large scale datasets created from user labels or openly available data have become crucial to provide training data for large scale learning algorithms. While these datasets are easier to acquire, the data are frequently noisy and unreliable, which is motivating research on weakly supervised learning techniques. In this paper we propose an iterative learning method that extracts the useful information from a large scale change detection dataset generated from open vector data to train a fully convolutional network which surpasses the performance obtained by naive supervised learning. We also propose the guided anisotropic diffusion algorithm, which improves semantic segmentation results using the input images as guides to perform edge preserving filtering, and is used in conjunction with the iterative training method to improve results.