Estimating Depth From Monocular Images as Classification Using Deep Fully Convolutional Residual Networks

• Published in: IEEE transactions on circuits and systems for video technology Vol. 28; no. 11; pp. 3174 - 3182
• Main Authors: Cao, Yuanzhouhan, Wu, Zifeng, Shen, Chunhua
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2018; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Bins; Classification; deep residual networks; depth estimation; Estimation; Image classification; Network architecture; Neural networks; Performance enhancement; Post-processing; Predictive models; Probability distribution; Scene analysis; Semantics; State of the art; Statistical analysis; Test procedures; Training
• ISSN: 1051-8215; 1558-2205
• DOI: 10.1109/TCSVT.2017.2740321

Depth estimation from single monocular images is a key component in scene understanding. Most existing algorithms formulate depth estimation as a regression problem due to the continuous property of depths. However, the depth value of input data can hardly be regressed exactly to the ground-truth value. In this paper, we propose to formulate depth estimation as a pixelwise classification task. Specifically, we first discretize the continuous ground-truth depths into several bins and label the bins according to their depth ranges. Then, we solve the depth estimation problem as classification by training a fully convolutional deep residual network. Compared with estimating the exact depth of a single point, it is easier to estimate its depth range. More importantly, by performing depth classification instead of regression, we can easily obtain the confidence of a depth prediction in the form of probability distribution. With this confidence, we can apply an information gain loss to make use of the predictions that are close to ground-truth during training, as well as fully-connected conditional random fields for post-processing to further improve the performance. We test our proposed method on both indoor and outdoor benchmark RGB-Depth datasets and achieve state-of-the-art performance.