Meeting 12/2/2015

[gzuidhof]

• Action: Look up research papers of feature detection
  • Adam Coates 2011 - fast shallow learning feature extraction of images (K-means clustering works best and is the most simple method, but the method also works with GM's (EM), RBM and AE's)
  • Rotation invariance
  • Alex Krizhevsky 2012 - deep learning approach using convolutional neural networks (state of the art)
• Meetings
  • Wednesdays after AppLab (13:45 or 15:45 depending on App-Lab practical)
  • Saturdays 13:00 skype
• Communication
  • Slack? Hipchat?
  • Github Issues
• Tools
  • Python 2.7 (install Anaconda)
  • TeX (ShareLatex)

[StevenReitsma]

See my thesis (download and open in Adobe PDF, don't open in the Chrome/Firefox pdf reader cause it messes with the figures) for an explanation of Coates' method (chapter 3) and rotation invariance (chapter 4). Figure 4 is especially useful if you're having trouble understandig Coates' paper.