Scribd
Upload
14 views12 pages

DL-UNIT-3

The document outlines various applications of Convolutional Neural Networks (CNNs), including image classification, object detection, semantic segmentation, and medical image analysis. It also describes the building blocks of CNNs, such as the input image, convolution layers, pooling layers, and fully connected layers, detailing their functions and characteristics. Additionally, it mentions specific CNN architectures like AlexNet and VGG16, and discusses the classification of ResNet and VGG.

Uploaded by

Prafull Ravindra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
14 views12 pages

DL-UNIT-3

The document outlines various applications of Convolutional Neural Networks (CNNs), including image classification, object detection, semantic segmentation, and medical image analysis. It also describes the building blocks of CNNs, such as the input image, convolution layers, pooling layers, and fully connected layers, detailing their functions and characteristics. Additionally, it mentions specific CNN architectures like AlexNet and VGG16, and discusses the classification of ResNet and VGG.

Uploaded by

Prafull Ravindra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

1. Identify various applications of CNN.

a) Image Classification: CNNs are commonly used for image classification


tasks, where the network is trained to assign a label or category to an
input image. This application is prevalent in areas such as object
recognition, face recognition, and scene classification.
b) Object Detection: CNNs excel in object detection, where the goal is to
identify and locate objects within an image. This is used in autonomous
vehicles, surveillance systems, and various other applications where
identifying specific objects in an image is crucial.
c) Semantic Segmentation: CNNs are employed for semantic segmentation,
which involves classifying each pixel in an image into specific classes. This
is useful in medical image analysis, autonomous navigation, and scene
understanding.
d) Face Recognition: CNNs have been widely used in face recognition
systems, enabling the identification and verification of individuals based on
facial features. This has applications in security, access control, and
personal device authentication.
e) Gesture Recognition: CNNs can be trained to recognize gestures and
movements, making them valuable in applications like sign language
interpretation, virtual reality, and human-computer interaction.
f) Medical Image Analysis: CNNs are applied to analyze medical images for
tasks such as tumor detection, organ segmentation, and disease diagnosis.
The ability of CNNs to learn hierarchical features from images is
particularly useful in this domain.
g) Video Analysis: CNNs can be extended to analyze video data. Applications
include action recognition, tracking objects or people in videos, and video
summarization.
h) Style Transfer: CNNs have been used in artistic applications, such as style
transfer, where the style of one image is applied to another. This is often
used to create visually appealing images and videos.
i) Natural Language Processing (NLP): While primarily designed for image-
related tasks, CNNs can also be applied in conjunction with other models
for certain NLP tasks, such as text classification and sentiment analysis, by
treating words as one-dimensional sequences.
j) Autonomous Vehicles: CNNs play a crucial role in the development of
perception systems for autonomous vehicles. They are used for tasks like
lane detection, object detection, and obstacle avoidance.
k) Satellite Image Analysis: CNNs are employed in the analysis of satellite
and aerial imagery for various purposes, including land cover classification,
disaster monitoring, and environmental analysis.
l) Quality Control in Manufacturing: CNNs can be used for inspecting and
ensuring the quality of manufactured products by analyzing images for
defects, irregularities, or deviations from standards.

2. Analyze building blocks of CNN.

Input Image

Greyscale VS RGB Image (Source: datahacker.rs)

First of all, the input image will be broken down into pixels. If it is a black
and white image, it will only have one layer and pixels will be interpreted as
2D array with the value from 0 to 255. If it is colored image, it will have 3
layers (red, green, blue) and will be interpreted as 3D array.
Convolution Layer

Convolution Layer (Source: indoml.com)

This is the first layer that filters the input images. Its purpose is to extract
features from the image. It captures color, edges, gradient orientation, and
other features so it can be differentiated.

A convolutional layer within a neural network should have the following


attributes:
· Convolutional kernels defined by a width and height (hyper-parameters).
· The number of input channels and output channels (hyper-parameter).
· The depth of the Convolution filter (the input channels) must be equal to
the number channels (depth) of the input feature map.

There are two types of result from this layer:


· Same Padding — When the size of output feature-maps are the same as the
input feature-maps
· Valid Padding — When the size of output feature-maps has the same size as
the kernel
Pooling Layer

Pooling Layer Illustration and Comparison (Source: Researchgate)

This layer is usually added after the convolutional layer. Pooling layer
reduces the spatial size of the output from convolutional layer and extracts
dominant features. Pooling layer can be differentiated into two types, which
are:
· Max Pooling
It returns the maximum value from the portion of the image covered by
kernel. This layer discards noisy activation and help over-fitting by providing
an abstracted form of the representation.
· Average Pooling
This type of pooling returns the average value from the portion of the image
covered by kernel.
There’s no way we can say max pooling is better than average pooling or vice
versa. The type of pooling we use depends on the input image that we “feed”
the model. From the image above, we can see that in case (a), it is better to
use average pooling but in case (b), max pooling is better.
Fully Connected Input Layer (Flatten)

Flattened layer (Source: superdatascience.com)

Fully connected layers are layers where all the inputs from one layer are
connected to every activation unit of the next layer. The layer takes the
output of the pooling and flatten them into single vector.

Fully Connected Layer

Fully Connected Layer (Source: Superdatascience)

The flattened matrix goes through a fully connected layer to classify the
images. The purpose of this layer is to classify the image into a label. It takes
the output of previous layer and predicts the best label by applying weights
and “voting”. The final output will be the probabilities for each label.
3. Describe convolutional neural network in detail.
4. Explain Alexnet in detail.
5. Describe VGG 16 in detail.
6. Summarize Multiple filters at one time in CNN in detail.
7. Classify ResNet and VGG.
8. Discuss any one convnet architecture in detail.

You might also like