CS-231n Winter 2016 ( Convolutional Neural Networks for Visual Recognition )

Cat

Have a look at Course-Description

#### Cs-231n encountered its second series in 2016 taught by Prof Lei Fei Fei and Andrej Karpathy at Stanford University.

Lecture 1 : Data-driven approach, kNN, Linear Classification

1. Python

First of all, understanding the process of building neural networks requires a basic understanding of Python. I have compiled some important fundamentals in Python_Basics.py to get started.

Useful Links to get started :

2. Image Classification- A core task in Computer Vision

Perception

An image is composed of a 2d matrix of numbers also called pixels . For a colored picture we have 3 layers of Red, Green, Blue 2d matrices stacked together one by one.

What is Data Driven Approach ?

  1. Collect a dataset of images and labels.
  2. Use Machine Learning to train an image classifier.
  3. Evaluate the classifier on a withheld set of images.

Lets start :

Classifier 1: K Nearest Neighbor

Dataset: CIFAR 10, Labels : 10, Training Images : 50000, Test Images : 10000

Lecture 2 : Linear Classification 2, Optimization