You are on page 1of 2

Introduction

In this project we will be training matlab with different image categories , and by using algorithm
classify the test image .

Aim and objective:

The aim of the project is to load, train and test different categories of images using deep learning in
matlab and to obtain result accuracy on test images using SVM classifier for scene classification.

To load different types of images in matlab

To train by extracting feature for scene classification

To test images for scene classification.

Summary OF LR

What is AI ?

Artificial intelligence is a technology where it is train to be smart as a human being. By focusing on


the aspects of how human brain think, learn and decide. The idea of machines operating like human
beings began to be the centre of the mind of the scientist, and the mathematician Alan Turing
introduced whether machines can be made with the same ability to think and learn by themselves. It
turns out after a series of tests (later called Turing Test) that machines can think and learn just like
humans. ARTIFICIAL INTELLIGENT IS CLASSIFIED into two main TYPES where as in the 1st type
narrow AI, general ai and super Ai and in the 2nd type reactive machine, limited memory , theory of
mind and self-awareness.

And there many braches of ai such as vision , robotics, speech, in which one of the main is machine
learning .

Machine learning is a technique where the machine learns itself by analysing the data, recognizing
the pattern and making the decision

There are 4 types of machine learning :

Supervised learning: where machine learn itself, the learning algorithm receive a set of inputs with
the appropriate outputs. Let’s take a basic example here, the appropriate outputs are pass or fail,
the set of inputs are the yearly data of a student, it analysis the data and make the decision whether
it is pass or fail.

Unsupervised learning: It is train on historical or pervious data. Let take an example here, you are
planning to buy a new phone and you started googling it . specification, colour, how much, where
to buy etc… it has learned from pervious data that u are going to buy phone. So after analysing the
data, it started to show irrelevant adds of phone .

Semi-supervised: It is used as supervised learning in same applications, But for training, it uses both
relevant and irrelevant data, due to it is not easy to acess relevant data as in in the example yearly
data of a student. So it combine both data and come up with a decision. Early example face on the
web cam.
REINFORCEMENT LEARNING: it is often used for robotics, gaming and navigation. The algorithm
discovers through trial and error with reinforcement learning which actions yield the greatest
reward

Deep learning is a subset of machine learning which trains the comes to people: learn by example.
Deep learning is a key technology behind driverless cars that makes it possible for them to recognize
a stop sign or distinguish a pedestrian from a lamp post. Deep learning can perform many tasks, such
as scene classifying, objects detection. But there is one disadvantage, it requires more data as 1000
to million of data to analyse and come up with a accurate outcome.

In this experiment, we are using bag of feature technique to extract feature from different images,
for example let take a scene of highway and beach, so what d u c on a highway, tress, road, sand sky
etc… and on a beach sand, tress, water, sky etc… as well as in auditorium and a restaurant, table,
chair, screen food etcc…

After extracting all this feature we are using svm classifier algorithm for scene classification.

You might also like