DepthFormer: Exploiting Long-Range Correlation and Local Information for Accurate Monocular Depth Estimation

TL;DR

DepthFormer leverages a hybrid Transformer-CNN architecture with hierarchical aggregation and deformable attention to improve monocular depth estimation accuracy, outperforming existing methods on multiple benchmarks.

Contribution

The paper introduces DepthFormer, a novel model combining global Transformer attention with local CNN features, and a hierarchical interaction module for enhanced depth estimation.

Findings

Outperforms state-of-the-art on KITTI, NYU, and SUN RGB-D datasets.

Achieves top results on the KITTI depth estimation benchmark.

Demonstrates the effectiveness of long-range correlation modeling in depth estimation.

Abstract

This paper aims to address the problem of supervised monocular depth estimation. We start with a meticulous pilot study to demonstrate that the long-range correlation is essential for accurate depth estimation. Therefore, we propose to leverage the Transformer to model this global context with an effective attention mechanism. We also adopt an additional convolution branch to preserve the local information as the Transformer lacks the spatial inductive bias in modeling such contents. However, independent branches lead to a shortage of connections between features. To bridge this gap, we design a hierarchical aggregation and heterogeneous interaction module to enhance the Transformer features via element-wise interaction and model the affinity between the Transformer and the CNN features in a set-to-set translation manner. Due to the unbearable memory cost caused by global attention on high-resolution feature maps, we introduce the deformable scheme to reduce the complexity. Extensive experiments on the KITTI, NYU, and SUN RGB-D datasets demonstrate that our proposed model, termed DepthFormer, surpasses state-of-the-art monocular depth estimation methods with prominent margins. Notably, it achieves the most competitive result on the highly competitive KITTI depth estimation benchmark. Our codes and models are available at https://github.com/zhyever/Monocular-Depth-Estimation-Toolbox.