Monday, December 9, 2019
  • Setup menu at Appearance » Menus and assign menu to Top Bar Navigation
Advertisement
  • AI Development
    • Artificial Intelligence
    • Machine Learning
    • Neural Networks
    • Learn to Code
  • Data
    • Blockchain
    • Big Data
    • Data Science
  • IT Security
    • Internet Privacy
    • Internet Security
  • Marketing
    • Digital Marketing
    • Marketing Technology
  • Technology Companies
  • Crypto News
No Result
View All Result
NikolaNews
  • AI Development
    • Artificial Intelligence
    • Machine Learning
    • Neural Networks
    • Learn to Code
  • Data
    • Blockchain
    • Big Data
    • Data Science
  • IT Security
    • Internet Privacy
    • Internet Security
  • Marketing
    • Digital Marketing
    • Marketing Technology
  • Technology Companies
  • Crypto News
No Result
View All Result
NikolaNews
No Result
View All Result
Home Neural Networks

Classifying Flowers with CNNS and Transfer Learning

December 3, 2019
in Neural Networks
Classifying Flowers with CNNS and Transfer Learning
585
SHARES
3.2k
VIEWS
Share on FacebookShare on Twitter

If I asked you what type of flower is pictured above, you would probably know it’s a sunflower. But what if I asked a computer the same question? Wouldn’t it be pretty impressive if it answered correctly? Remarkably, computers can be trained to classify not only sunflowers but a multitude of other flowers as well. And the person training the computer doesn’t even have to know what a sunflower is!

To identify types of flowers, I developed a Convolutional Neural Network (CNN) that can classify dandelions, daisies, tulips, sunflowers, and roses. Check out the full code HERE

You might also like

Machine Learning Biology Roots – Becoming Human: Artificial Intelligence Magazine

GANs: How to make things that aren’t there

Machine Learning vs. Deep Learning vs. Data Science

What are neural networks? Artificial neural networks are modeled after the biological neural networks of the human brain and can make predictions based on patterns recognized in past data. These networks have three types of layers: an input layer where initial data is provided, a hidden layer where weights and biases are used to perform computations, and an output layer where an activation function is applied to give the final results. Read a more in-depth description of neural networks here.

A neural network visualized

A convolutional neural network is a type of neural network that typically includes convolutional layers and max-pooling layers. A convolution is simply applying a filter onto the collection of pixels that make up the input image. This results in an activation. Repeated applications of this filter result in a map of activations called a feature map which essentially tells the computer about an image. Following the convolutional layer is the max-pooling layer. In the max-pooling layer, the filter on the image checks for the greatest pixel value in each section (the size of the section is specified by the programmer) and then uses the maximum pixel values to create a new, smaller image. These smaller images help the computer run the model much faster. Check this video out for a deeper description of CNNs

When the convolutional layers and max-pooling layers are connected to the input and output layers of a neural network, the model is able to use past labeled data to make predictions of what future images contain. Now that we understand what a CNN is, let’s look at the steps to build one.

We can code this project using Python and the TensorFlow library. The flowers dataset (containing labeled images of the 5 classes of flowers) is already provided in TensorFlow Datasets so it can simply be downloaded from there. Yet, the dataset must be refined before it can be passed in the model. We split 70% of the data into the training set and the remaining 30% of the data into the validation set. The training set is a set of examples used by the model to fit the parameters of the classifier and the validation set is used to further tune these parameters so that the classifier can work on data not seen before. Since the images of this dataset are of different sizes, we resize all of them to a standard size.

After gathering the data, we can begin to train our model. Transfer learning is the process of reusing parts of an already trained model and changing the final layer of the model; the final layer is then retrained on the flowers dataset to give the outputs we want. Transfer learning is implemented because it can improve the accuracy of our model. Using the MobileNet v2 model and changing only the final layer makes the code for the actual model very short.

After training the model, we can plot its accuracy and loss to see how it is doing. The x-axis on the graph represents the number of epochs (number of times the model runs through the entire training set and updates the weights).

The model reaches a 90% accuracy on the validation set!

Let’s look at the results after we run the image batch through the model!

As you can see, the model performed very well on the testing set as a result of transfer learning and the CNN architecture.

So how exactly is this useful? While classifying flowers may only be helpful to botanists, CNNs can have life-saving applications such as detecting pneumonia from MRIs and making self-driving cars a reality.

Don’t leave yet!

I’m Roshan, a 16 year old passionate about technologies such as Artifical Intelligence/Machine Learning and the Internet of Things.

If you enjoyed this article or have an interest in disruptive technology, make sure to:

Credit: BecomingHuman By: Roshan Adusumilli

Previous Post

This trojan malware is being used to steal passwords and spread ransomware

Next Post

What to Ask When Implementing Machine Learning 

Related Posts

Machine Learning Biology Roots – Becoming Human: Artificial Intelligence Magazine
Neural Networks

Machine Learning Biology Roots – Becoming Human: Artificial Intelligence Magazine

December 7, 2019
GANs: How to make things that aren’t there
Neural Networks

GANs: How to make things that aren’t there

December 7, 2019
Machine Learning vs. Deep Learning vs. Data Science
Neural Networks

Machine Learning vs. Deep Learning vs. Data Science

December 7, 2019
3 travel tech predictions for 2020
Neural Networks

3 travel tech predictions for 2020

December 6, 2019
7 Reasons Why A Smartwatch Can Be Beneficial For Your Health With AI
Neural Networks

7 Reasons Why A Smartwatch Can Be Beneficial For Your Health With AI

December 6, 2019
Next Post
What to Ask When Implementing Machine Learning 

What to Ask When Implementing Machine Learning 

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Recommended

Plasticity in Deep Learning: Dynamic Adaptations for AI Self-Driving Cars

Plasticity in Deep Learning: Dynamic Adaptations for AI Self-Driving Cars

January 6, 2019
Microsoft, Google Use Artificial Intelligence to Fight Hackers

Microsoft, Google Use Artificial Intelligence to Fight Hackers

January 6, 2019

Categories

  • Artificial Intelligence
  • Big Data
  • Blockchain
  • Crypto News
  • Data Science
  • Digital Marketing
  • Internet Privacy
  • Internet Security
  • Learn to Code
  • Machine Learning
  • Marketing Technology
  • Neural Networks
  • Technology Companies

Don't miss it

Qeexo AutoML Demo: Automating Machine Learning for Embedded Devices
Machine Learning

Qeexo AutoML Demo: Automating Machine Learning for Embedded Devices

December 9, 2019
Discover how machine learning can solve finance industry challenges by Jannes Klaas
Data Science

Event Distribution as a Subject of Ontological Recognition Criteria

December 9, 2019
Why the Trail Blazers’ NBA Playoff Hopes May Be Completely Doomed
Crypto News

Why the Trail Blazers’ NBA Playoff Hopes May Be Completely Doomed

December 9, 2019
AWS SageMaker’s new machine learning IDE isn’t ready to win over data scientists
Machine Learning

AWS SageMaker’s new machine learning IDE isn’t ready to win over data scientists

December 9, 2019
US charges two members of the Dridex malware gang
Internet Security

US charges two members of the Dridex malware gang

December 9, 2019
Learn Python for data science and machine learning for just $10
Machine Learning

Learn Python for data science and machine learning for just $10

December 8, 2019
NikolaNews

NikolaNews.com is an online News Portal which aims to share news about blockchain, AI, Big Data, and Data Privacy and more!

What’s New Here?

  • Qeexo AutoML Demo: Automating Machine Learning for Embedded Devices December 9, 2019
  • Event Distribution as a Subject of Ontological Recognition Criteria December 9, 2019
  • Why the Trail Blazers’ NBA Playoff Hopes May Be Completely Doomed December 9, 2019
  • AWS SageMaker’s new machine learning IDE isn’t ready to win over data scientists December 9, 2019

Subscribe to get more!

© 2019 NikolaNews.com - Global Tech Updates

No Result
View All Result
  • AI Development
    • Artificial Intelligence
    • Machine Learning
    • Neural Networks
    • Learn to Code
  • Data
    • Blockchain
    • Big Data
    • Data Science
  • IT Security
    • Internet Privacy
    • Internet Security
  • Marketing
    • Digital Marketing
    • Marketing Technology
  • Technology Companies
  • Crypto News

© 2019 NikolaNews.com - Global Tech Updates