Course assignment for Image Analysis and Understanding 2017

A python implementation of depth estimation&plane detection described in the papers:

• Accurate 3D ground plane estimation from a single image(ICRA 2009)
• Depth Map Prediction from a Single Image using a Multi-Scale Deep Network.(Nips 2014)

the depth-estimation algorithm use CNN in the second paper and the plane detection algorithm is in the first paper

1. Python 2.7
2. Python packages : numpy, scipy, PIL(or Pillow), opencv,caffe for python, django
3. Pretrained caffe model : model_norm_abs_100k.caffemodel

      * Please download the file from link above.
      * Save the caffemodel file under /model

python manage.py runserver

• The following figure shows the screenshot of our demo website.

• These are the depth estimation and plane detection figure.

My implementation is based on the projects:

• This is a implemention of depth-estimation using CNN, I use the pretrained caffe model in that project.

• It is our project website.

• Spring 2017: Image Analysis and Understanding(Our course website)

This implementation has been tested with python2.7 and Ubuntu 16.04.

My core code to process pictures are in homework/depth.py and homework/plane.py with comments.