Depth Extraction from Videos Using Geometric Context and Occlusion Boundaries

TL;DR

This paper introduces a novel method for estimating depth in dynamic videos by leveraging appearance, motion, occlusion boundaries, and geometric context without requiring camera pose information, validated on a new dataset.

Contribution

The authors develop a depth estimation algorithm that works on unconstrained videos using a combination of learned appearance-depth relationships and occlusion cues within an MRF framework, without camera pose estimation.

Findings

Accurate depth maps for outdoor videos with background and foreground motion

Effective temporal smoothing for consistent depth estimation

Superior performance on new and existing datasets

Abstract

We present an algorithm to estimate depth in dynamic video scenes. We propose to learn and infer depth in videos from appearance, motion, occlusion boundaries, and geometric context of the scene. Using our method, depth can be estimated from unconstrained videos with no requirement of camera pose estimation, and with significant background/foreground motions. We start by decomposing a video into spatio-temporal regions. For each spatio-temporal region, we learn the relationship of depth to visual appearance, motion, and geometric classes. Then we infer the depth information of new scenes using piecewise planar parametrization estimated within a Markov random field (MRF) framework by combining appearance to depth learned mappings and occlusion boundary guided smoothness constraints. Subsequently, we perform temporal smoothing to obtain temporally consistent depth maps. To evaluate our depth estimation algorithm, we provide a novel dataset with ground truth depth for outdoor video scenes. We present a thorough evaluation of our algorithm on our new dataset and the publicly available Make3d static image dataset.