AutoCorrect: Deep Inductive Alignment of Noisy Geometric Annotations
This addresses the issue of noisy annotations in computer vision, particularly for geo-spatial imagery, with incremental improvements in annotation correction methods.
The paper tackles the problem of correcting noisy geometric annotations in datasets by proposing AutoCorrect, a method that uses a consistency loss to train deep neural networks to align and correct annotations, achieving new state-of-the-art results on benchmarks like INRIA Buildings.
We propose AutoCorrect, a method to automatically learn object-annotation alignments from a dataset with annotations affected by geometric noise. The method is based on a consistency loss that enables deep neural networks to be trained, given only noisy annotations as input, to correct the annotations. When some noise-free annotations are available, we show that the consistency loss reduces to a stricter self-supervised loss. We also show that the method can implicitly leverage object symmetries to reduce the ambiguity arising in correcting noisy annotations. When multiple object-annotation pairs are present in an image, we introduce a spatial memory map that allows the network to correct annotations sequentially, one at a time, while accounting for all other annotations in the image and corrections performed so far. Through ablation, we show the benefit of these contributions, demonstrating excellent results on geo-spatial imagery. Specifically, we show results using a new Railway tracks dataset as well as the public INRIA Buildings benchmarks, achieving new state-of-the-art results for the latter.