[[nips_3d_pose:main|NIPS 3D pose estimation]]

====== How to test an existing model ======

===== test_caffe_model.sh =====

This script runs a trained model on a given test-set, and outputs absolute error (mm), transformed (mm) and pixel error (on 220x220 images).

  inputs:
    - caffemodel trained
    - clustering used during training (matlab file)
    - network definition file (DEPLOY version of trained network)
    - test set (.h5 file, must contain '/data' and '/joints3D_centered')

  dependencies/called functions:
    - source_common_deps.sh : Caffe path, Matlab path, etc..
    - extract_data_h5.m : extracts images from an .h5 file
    - caffe_for_matlab_ubuntu.py : runs Caffe from a given set of images in Matlab .mat format
    - compute_absolute_error.m : given a testfile, a clustering, and the caffe output, compute absolute error

===== Preparing input =====

  - h5 test set: joints3D_centered should have joints centered around the back of the skeleton, in meters