Single Image Depth Estimation: An Overview

TL;DR

This paper provides a comprehensive overview of single image depth estimation, highlighting the evolution from traditional methods to modern deep learning techniques, and discussing various approaches, challenges, and future directions.

Contribution

It offers a detailed survey of both non-deep learning and deep learning methods, including supervised, unsupervised, and multitask approaches, for single image depth estimation.

Findings

Deep learning methods dominate current solutions.

Supervised and unsupervised approaches each have unique advantages.

Multitask learning improves depth estimation accuracy.

Abstract

We review solutions to the problem of depth estimation, arguably the most important subtask in scene understanding. We focus on the single image depth estimation problem. Due to its properties, the single image depth estimation problem is currently best tackled with machine learning methods, most successfully with convolutional neural networks. We provide an overview of the field by examining key works. We examine non-deep learning approaches that mostly predate deep learning and utilize hand-crafted features and assumptions, and more recent works that mostly use deep learning techniques. The single image depth estimation problem is tackled first in a supervised fashion with absolute or relative depth information acquired from human or sensor-labeled data, or in an unsupervised way using unlabelled stereo images or video datasets. We also study multitask approaches that combine the depth estimation problem with related tasks such as semantic segmentation and surface normal estimation. Finally, we discuss investigations into the mechanisms, principles, and failure cases of contemporary solutions.