This paper addresses the problem of depth estimation from a single stillimage. Inspired by recent works on multi- scale convolutional neural networks(CNN), we propose a deep model which fuses complementary information derivedfrom multiple CNN side outputs. Different from previous methods, theintegration is obtained by means of continuous Conditional Random Fields(CRFs). In particular, we propose two different variations, one based on acascade of multiple CRFs, the other on a unified graphical model. By designinga novel CNN implementation of mean-field updates for continuous CRFs, we showthat both proposed models can be regarded as sequential deep networks and thattraining can be performed end-to-end. Through extensive experimental evaluationwe demonstrate the effective- ness of the proposed approach and establish newstate of the art results on publicly available datasets.
展开▼
