CS231n

Lecture 1. Intoroduction to Convolutional Neural Networks for Visual Recognition

History of Computer vision

Input image ~ Edge image ~ 2 1/2D sketch ~ 3D model

how can we move?

how we can recognize real workd object?

if object recognitio is too hard, maybe we should first do object segmentation,

that is the task of taking an image and goup the fixels into meaningful area.

• Normalized Cut

face recognition

간단하게 역사 설명만 했다.

CS231 Overview

Image classification ex) recognize food,

Action classification

image captioning

Convolutional Neural Networks

1. computing available

2. data - before: algorithms are hungry for data

   ex) pascal, dataset