NEW related work!!

Cross-Spectral Local Descriptors via Quadruplet Network

Learning cross-spectral similarity measures with deep convolutional neural networks

Bibtex

@inproceedings{Aguilera_cvprw_2016,
    organization = { IEEE  },
    year = { 2016  },
    pages = { 9  },
    month = { Jun  },
    location = { Las vegas, USA  },
    booktitle = { The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops  },
    author = { Cristhian A. Aguilera and Francisco J. Aguilera and Angel D. Sappa and Cristhian Aguilera and Ricardo Toledo  },
    title = { Learning cross-spectral similarity measures with deep convolutional neural networks  },
}

Instructions

First install the torch framework and cudnn

Datasets

Nirscenes patches

Follow the following steps to generate the dataset

Download gt csv
```
cd datasets
./download_nirscenes_csv.sh
```
Download the original nirscenes dataset link
Decompress the dataset on /datasets/nirscenes
Convert the images in each folder to ppm (Torch doesn't support tiff). Use your favorite software
Install csvigo
```
luarocks install csvigo
```
Use our script to generate the dataset
```
cd utils
th nirscenes_to_t7.lua
```

The t7 dataset is stored in datasets/nirscenes

VIS-LWIR ICIP2015

Download the dataset

cd datasets
./download_icip_dataset.sh

It's easier if you are the owner of the dataset :)

Eval

Nirscenes eval (cpu and cuda support)

Evaluation code can be found in the eval folder. To eval one sequence:

You have to generate the nirscenes patch dataset
Install xlua
```
luarocks install xlua
```

Run

cd eval
th nirscenes_eval.lua -seq_path ../dataset/nirscenes/[sequence].t7 -net .. [trained network]

For example, to eval the field sequence using the 2ch_country network bash th nirscenes_eval.lua -seq_path ../dataset/nirscenes/field.t7 -net ../trained_networks/2ch_country.net -net_type 2ch

For more options, run bash th nirscenes_eval -h

VIS-LWIR eval (ICIP2015) (just cuda support)

You have to download the dataset first

Run

cd eval
th icip2015_eval.lua -dataset_path ../nirscenes/icip2015/ -net [trained network]

For example. To eval 2ch_country

```bash
cd eval
th icip2015_eval.lua -dataset_path ../nirscenes/icip2015/ -net ../trained_networks/2ch_country.t7
```

Training

Install penlight
```
luarocks install penlight
```

Train a network

cd train
th nirscenes_doall.lua -training_sequences [country|field|...] -net [2ch|siam|psiam]

For example, train a 2ch network using the country sequence

cd train
th nirscenes_doall.lua -training_sequences country -net 2ch

Results will be stored in the results folder.For more options, run

th nirscenes_doall.lua -h

Note The training code is different from the one used in the article. This new version is much faster.

whuwhu123 / lcsis