This is a python project of deep learning for scene labeling.

Course Project in Telecom Paristech

---

Here is the structure of the project,

prim_dl
├── data
│   ├── data_stanford
│   │   ├── images
│   │   └── labels
│   └── test
└── scripts

• data, data forlder, ignored by git. We train the model by the backgroud dataset from Stanford Lab. Also there are data below test for debug
• scripts, contain the relative fonctions

---

Environment

I write this instruction under ubuntu 14.04

• Install Anaconda from https://www.continuum.io/downloads
• Install Theano from http://deeplearning.net/software/theano/install.html#anaconda
• Install TensorFlow from https://www.tensorflow.org/versions/r0.8/get_started/os_setup.html
• Install OpenCV from https://anaconda.org/search?q=opencv
• Install C++ by sudo apt-get install build-essential
• Install imagemagick for PIL by sudo apt-get install imagemagick

---

Fonctions

file_system.py

• Create new folder
• Delete the folder with subfolder

image_proc.py

• Open and convert the image to a special type
• Show the image
• Get the laplacien pyramid
• Lecun local contrast normalization

cnn

• To be done ...

segmentation

• To be done ...