Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
94 views

Deep Learning For Image Processing Using MATLAB

This document discusses deep learning for image processing using MATLAB. It covers types of convolutional neural networks (CNNs), including training CNNs from scratch and using pre-trained CNNs for transfer learning. It explains that deep learning is preferred over machine learning for image processing tasks. It also describes using pre-trained networks that have been trained on large datasets for transfer learning to a new problem, and provides examples of training a CNN from scratch and fine-tuning a pre-trained ResNet-18 model for concrete fault classification.

Uploaded by

Netaji Gandi
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
94 views

Deep Learning For Image Processing Using MATLAB

This document discusses deep learning for image processing using MATLAB. It covers types of convolutional neural networks (CNNs), including training CNNs from scratch and using pre-trained CNNs for transfer learning. It explains that deep learning is preferred over machine learning for image processing tasks. It also describes using pre-trained networks that have been trained on large datasets for transfer learning to a new problem, and provides examples of training a CNN from scratch and fine-tuning a pre-trained ResNet-18 model for concrete fault classification.

Uploaded by

Netaji Gandi
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 19

Deep learning for Image

Processing using MATLAB


Ms. Manisha Saini
Assistant Professor
G D Goenka University
Types of CNN:-
1) Convolutional Neural Networks (CNN) trained from scratch
2) Pre-trained convolutional neural networks (Transfer learning)
Why Deep learning is preferred
over Machine learning ??
Pre-trained Networks

By using pre-trained models which have been previously trained on


large datasets, we can directly use the weights and architecture
obtained and apply the learning on our problem statement. This is
known as transfer learning. We are transfer the learning of the pre-
trained model to our specific problem statement.

Simply put, a pre-trained model is a model created by some one else


to solve a similar problem. Instead of building a model from scratch to
solve a similar problem, you use the model trained on other problem
as a starting point.
Convolutional Network(CNN)
MATLAB Implementation
(using Deep Learning Toolbox)
Training CNN on Concrete fault classification
subset dataset
Prediction using the Pre-trained
Network(ResNet-18)
Training RESNET-18 transfer Learning on
Concrete fault classification subset dataset

You might also like