GA-Net: Guided Aggregation Net for End-to-end Stereo Matching

TL;DR

GA-Net introduces two novel neural network layers for cost aggregation in stereo matching, replacing costly 3D convolutions, leading to improved accuracy and efficiency on benchmark datasets.

Contribution

The paper proposes semi-global and local guided aggregation layers that enhance stereo matching accuracy while reducing computational complexity.

Findings

Outperforms state-of-the-art GC-Net with fewer layers

Achieves better accuracy on Scene Flow and KITTI datasets

Reduces computational cost compared to 3D convolutional approaches

Abstract

In the stereo matching task, matching cost aggregation is crucial in both traditional methods and deep neural network models in order to accurately estimate disparities. We propose two novel neural net layers, aimed at capturing local and the whole-image cost dependencies respectively. The first is a semi-global aggregation layer which is a differentiable approximation of the semi-global matching, the second is the local guided aggregation layer which follows a traditional cost filtering strategy to refine thin structures. These two layers can be used to replace the widely used 3D convolutional layer which is computationally costly and memory-consuming as it has cubic computational/memory complexity. In the experiments, we show that nets with a two-layer guided aggregation block easily outperform the state-of-the-art GC-Net which has nineteen 3D convolutional layers. We also train a deep guided aggregation network (GA-Net) which gets better accuracies than state-of-the-art methods on both Scene Flow dataset and KITTI benchmarks.