Multi-view stereo infrared dataset. URL: https://zefirus.org/datasets/MVSIR.zip
3D models of three test objects used in the MVSIR dataset in Stanford .PLY format can be found in the models
folder.
- Copy
code
directory to~\path-to-caffe\examples\
- Create an LMDB with 28x28 grayscale images and specify the path in
MVS_IR_016.prototxt
- Use
MVS_IR_016.sh
to train the network
If you use this code for your research, please cite our paper: Deep Learning of Convolutional Auto-Encoder for Image Matching and 3D Object Reconstruction in the Infrared Range
@InProceedings{Knyaz_2017_ICCV,
author = {Knyaz, Vladimir A. and Vygolov, Oleg and Kniaz, Vladimir V. and Vizilter, Yury and Gorbatsevich, Vladimir and Luhmann, Thomas and Conen, Niklas},
title = {Deep Learning of Convolutional Auto-Encoder for Image Matching and 3D Object Reconstruction in the Infrared Range},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {Oct},
year = {2017}
}