Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue

TL;DR

This paper introduces an unsupervised CNN framework for single view depth estimation that learns from stereo pairs without manual labels or depth annotations, achieving competitive results.

Contribution

It presents a novel unsupervised training method for depth prediction using photometric consistency from stereo pairs, eliminating the need for ground truth depths or pre-training.

Findings

Achieves comparable performance to supervised methods on KITTI dataset.

Requires less than half of the KITTI data for training.

No manual annotation or depth sensor calibration needed.

Abstract

A significant weakness of most current deep Convolutional Neural Networks is the need to train them using vast amounts of manu- ally labelled data. In this work we propose a unsupervised framework to learn a deep convolutional neural network for single view depth predic- tion, without requiring a pre-training stage or annotated ground truth depths. We achieve this by training the network in a manner analogous to an autoencoder. At training time we consider a pair of images, source and target, with small, known camera motion between the two such as a stereo pair. We train the convolutional encoder for the task of predicting the depth map for the source image. To do so, we explicitly generate an inverse warp of the target image using the predicted depth and known inter-view displacement, to reconstruct the source image; the photomet- ric error in the reconstruction is the reconstruction loss for the encoder. The acquisition of this training data is considerably simpler than for equivalent systems, requiring no manual annotation, nor calibration of depth sensor to camera. We show that our network trained on less than half of the KITTI dataset (without any further augmentation) gives com- parable performance to that of the state of art supervised methods for single view depth estimation.