[docker] add support for docker

Dockerfile

@@ -0,0 +1,69 @@
+ARG CUDA_TAG=9.2-devel
+FROM nvidia/cuda:$CUDA_TAG
+LABEL maintainer="AliceVision Team alicevision@googlegroups.com"
+
+# use CUDA_TAG to select the image version to use
+# see https://hub.docker.com/r/nvidia/cuda/
+#
+# For example to create a ubuntu 16.04 with cuda 8.0 for development, use
+# CUDA_TAG=8.0-devel
+# docker build --build-arg CUDA_TAG=$CUDA_TAG --tag cctag:$CUDA_TAG .
+#
+# then execute with nvidia docker (https://github.com/nvidia/nvidia-docker/wiki/Installation-(version-2.0))
+# docker run -it --runtime=nvidia cctag
+
+# OS/Version (FILE): cat /etc/issue.net
+# Cuda version (ENV): $CUDA_VERSION
+
+# System update
+RUN apt-get clean && apt-get update && apt-get install -y --no-install-recommends \
+                build-essential \
+                cmake \
+                git \
+                wget \
+                unzip \
+                yasm \
+                pkg-config \
+                libtool \
+                nasm \
+                automake \
+                libpng12-dev \
+                libjpeg-turbo8-dev \
+                libboost-all-dev \
+                ffmpeg \
+                libavcodec-dev \
+                libavformat-dev \
+                libswscale-dev \
+                libeigen3-dev \
+                libavresample-dev \
+                libtbb-dev \
+        && rm -rf /var/lib/apt/lists/*
+
+WORKDIR /opt
+ENV OPENCV_VERSION="3.4.1"
+RUN wget https://github.com/opencv/opencv/archive/"${OPENCV_VERSION}".zip \
+&& unzip ${OPENCV_VERSION}.zip 
+
+WORKDIR /opt/opencv-${OPENCV_VERSION}/cmake_binary \
+RUN cmake -j -DBUILD_TIFF=ON \
+      -DBUILD_opencv_java=OFF \
+      -DWITH_CUDA=OFF \
+      -DENABLE_AVX=ON \
+      -DWITH_OPENGL=ON \
+      -DWITH_IPP=ON \
+      -DWITH_TBB=ON \
+      -DWITH_EIGEN=ON \
+      -DWITH_V4L=ON \
+      -DBUILD_TESTS=OFF \
+      -DBUILD_PERF_TESTS=OFF \
+      -DCMAKE_BUILD_TYPE=RELEASE  .. \
+    && make -j install \
+    && rm /opt/${OPENCV_VERSION}.zip \
+    && rm -r /opt/opencv-${OPENCV_VERSION}
+
+
+COPY . /opt/cctag
+WORKDIR /opt/cctag/build
+RUN cmake .. -DWITH_CUDA:BOOL=ON \
+       -DCMAKE_BUILD_TYPE=Release \
+       -DOpenCV_DIR:PATH=/usr/local/share/OpenCV && make install -j

BUILD.md → INSTALL.md

@@ -1,9 +1,7 @@
-CCTag library
-===================
+# CCTag library
+
+## Building instructions
 
-----------------------
-Building instructions
-----------------------
 
 Required tools:
 * CMake >= 3.4 to build the code
@@ -81,3 +79,22 @@ Several CUB functions are known to fail with a few NVidia cards including our re
 the GTX 980 Ti.
 The CUB that is included with CUDA 7.5 does not solve this problem.
 
+----------
+
+## Docker Image
+
+A docker image can be built using the Ubuntu based [Dockerfile](Dockerfile),which is based on nvidia/cuda image (https://hub.docker.com/r/nvidia/cuda/)
+
+A parameter `CUDA_TAG` can be passed when building the image to select the ubuntu and cuda version. 
+For example to create a ubuntu 16.04 with cuda 8.0 for development, use
+```
+docker build --build-arg CUDA_TAG=8.0-devel --tag cctag .
+```
+
+The complete list of available tags can be found on the nvidia [dockerhub page](https://hub.docker.com/r/nvidia/cuda/)
+In order to run the image nvidia docker is needed: see the installation instruction here https://github.com/nvidia/nvidia-docker/wiki/Installation-(version-2.0)
+Once installed, the docker can be run, e.g., in interactive mode with
+
+```
+docker run -it --runtime=nvidia cctag
+```

README.md

@@ -24,7 +24,7 @@ Check your graphic card CUDA compatibility [here](https://github.com/tpruvot/ccm
 Building
 --------
 
-See [BUILD](BUILD.md) text file.
+See [INSTALL](INSTALL.md) text file.
 
 Continuous integration:
  - [![Build Status](https://travis-ci.org/alicevision/CCTag.svg?branch=master)](https://travis-ci.org/alicevision/CCTag) master branch.