DeepMVS: Learning Multi-view Stereopsis
This addresses 3D reconstruction from images, a key problem in computer vision, with incremental improvements in handling challenging scenes.
The paper tackled multi-view stereo reconstruction by proposing DeepMVS, a deep convolutional neural network that predicts high-quality disparity maps from posed images, showing favorable results against state-of-the-art methods on the ETH3D Benchmark, especially for near-textureless regions and thin structures.
We present DeepMVS, a deep convolutional neural network (ConvNet) for multi-view stereo reconstruction. Taking an arbitrary number of posed images as input, we first produce a set of plane-sweep volumes and use the proposed DeepMVS network to predict high-quality disparity maps. The key contributions that enable these results are (1) supervised pretraining on a photorealistic synthetic dataset, (2) an effective method for aggregating information across a set of unordered images, and (3) integrating multi-layer feature activations from the pre-trained VGG-19 network. We validate the efficacy of DeepMVS using the ETH3D Benchmark. Our results show that DeepMVS compares favorably against state-of-the-art conventional MVS algorithms and other ConvNet based methods, particularly for near-textureless regions and thin structures.