Learning Depth from Single Images with Deep Neural Network Embedding Focal Length

TL;DR

This paper investigates the impact of focal length on monocular depth estimation, demonstrating that embedding focal length information into deep neural networks significantly improves depth recovery accuracy from single images.

Contribution

It introduces a method to generate synthetic varying-focal-length datasets and a neural network that effectively fuses middle-level features, outperforming existing methods.

Findings

Embedding focal length improves depth estimation accuracy.

Synthetic datasets enable training models with varying focal lengths.

Proposed method outperforms state-of-the-art on benchmark datasets.

Abstract

Learning depth from a single image, as an important issue in scene understanding, has attracted a lot of attention in the past decade. The accuracy of the depth estimation has been improved from conditional Markov random fields, non-parametric methods, to deep convolutional neural networks most recently. However, there exist inherent ambiguities in recovering 3D from a single 2D image. In this paper, we first prove the ambiguity between the focal length and monocular depth learning, and verify the result using experiments, showing that the focal length has a great influence on accurate depth recovery. In order to learn monocular depth by embedding the focal length, we propose a method to generate synthetic varying-focal-length dataset from fixed-focal-length datasets, and a simple and effective method is implemented to fill the holes in the newly generated images. For the sake of accurate depth recovery, we propose a novel deep neural network to infer depth through effectively fusing the middle-level information on the fixed-focal-length dataset, which outperforms the state-of-the-art methods built on pre-trained VGG. Furthermore, the newly generated varying-focal-length dataset is taken as input to the proposed network in both learning and inference phases. Extensive experiments on the fixed- and varying-focal-length datasets demonstrate that the learned monocular depth with embedded focal length is significantly improved compared to that without embedding the focal length information.