1. History of Computer Vision

  • Camera Obscura, 16th century
  • Hubel & Wiesel, 1959: visual processing mechanism with cats
  • Block World, Larry Roberts, 1963: geometric simplification of objects
  • MIT “The Summer Vision Project”, 1966: recogniztion with computers
  • Dabid Marr, late 1970s: Input -> Primal Sketch -> 2 1/2-D sketch -> 3-D Model Representation
  • Pictorial Structure, 1973
  • Generalized Cylinder, 1979
  • David Lowe, 1987
  • Normalized Cut, 1997: image segmentation
  • Dabid Lowe, “SITT & Object Recognition”, 1999: invariant features (SIFT feature)
  • Face Detection, 2001
  • Spatial Pyramid Matching, 2006: use features
  • Histogram of gradients, 2005: use features
  • Deformable part Model, 2009: using bounding boxes
  • PASCAL Visual Object Challenge, 2006~2012: performances are getting better
  • ImageNet
    • contribute to algorithm improvement
    • CNN, 2012

2. CS231n Overview

  • Image classification, object detection, and image captioning will be covered
  • Focus on CNN
    • Lecun et al, recognize digits, 1998 -> used the CNN concept for the first time
    • AlexNet, 2012 -> develops deep learning and increases layers
    • Advances in CNN: Improved hardwares and increased data contributed greatly



This is written by me after taking CS231n Spring 2017 provided by Stanford University. If you have questions, you can leave a reply on this post.