Structured Attention Guided Convolutional Neural Fields for Monocular Depth Estimation

TLDR

Integrating CRFs into deep networks improves pixel‑level predictions. The paper proposes a new method for monocular depth estimation. The method fuses multi‑scale CNN features with a continuous CRF enhanced by a structured attention module that controls inter‑scale information flow, enabling end‑to‑end training. Experiments show the approach matches prior methods on KITTI and surpasses the state of the art on NYU Depth V2.

Abstract

Recent works have shown the benefit of integrating Conditional Random Fields (CRFs) models into deep architectures for improving pixel-level prediction tasks. Following this line of research, in this paper we introduce a novel approach for monocular depth estimation. Similarly to previous works, our method employs a continuous CRF to fuse multi-scale information derived from different layers of a front-end Convolutional Neural Network (CNN). Differently from past works, our approach benefits from a structured attention model which automatically regulates the amount of information transferred between corresponding features at different scales. Importantly, the proposed attention model is seamlessly integrated into the CRF, allowing end-to-end training of the entire architecture. Our extensive experimental evaluation demonstrates the effectiveness of the proposed method which is competitive with previous methods on the KITTI benchmark and outperforms the state of the art on the NYU Depth V2 dataset.