Convolution Neural Network

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 4

Course: CONVOLUTION NEURAL NETWORK

Code: XXXXXX
4 Credits | Semester VI

A. Introduction: Convolutional Neural Networks (CNNs) are a powerful class of deep


learning models widely used in computer vision tasks such as image classification,
object detection, and image segmentation. This course provides CS&IT students with
a comprehensive understanding of CNNs, including their architecture, training
procedures, and applications. Topics covered include convolutional layers, pooling
layers, activation functions, CNN architectures (e.g., LeNet, AlexNet, VGG, ResNet),
transfer learning, and practical implementation using deep learning frameworks.
Through lectures, hands-on programming assignments, and projects, students will
gain the skills necessary to design, train, and evaluate CNN models for various
computer vision tasks.

B. Course Outcomes: At the end of the course, students will be able to:

[CO.1]. Understand the fundamental concepts and principles of Convolutional Neural Networks
(CNNs).
[CO.2]. Learn the architecture and components of CNNs, including convolutional layers, pooling
layers, and activation functions.
[CO.3]. Gain proficiency in building, training, and evaluating CNN models using deep learning
frameworks.
[CO.4]. Explore advanced topics in CNNs, such as transfer learning, data augmentation, and model
optimization.
[CO.5]. Apply CNNs to solve real-world computer vision problems and analyse model
performance.

C. Assessment Plan:

Criteria Description Maximum Marks


Internal Internal Examination 20
Assessment Assignment 05
(Summative)
Attendance 05
End Term Exam End Term Examination 70
(Summative)
Total 100
Attendance A minimum of 75% Attendance is required to be maintained by a student to
(Formative) be qualified for taking up the End Semester examination. The allowance of
25% includes all types of leaves including medical leaves.

D. SYLLABUS

Module 1: Introduction to Convolutional Neural Networks (CNNs)


Fundamentals of Neural Networks and Deep Learning, Introduction to Convolutional Neural
Networks (CNNs), Convolutional Layers: Filters, Stride, Padding, Pooling Layers: Max
Pooling, Average Pooling, Activation Functions: ReLU, Sigmoid, Tanh, CNN Architectures:
LeNet, AlexNet, VGG, ResNet

Module 2: Training and Optimization of CNNs


Training CNNs: Backpropagation and Optimization Algorithms, Regularization Techniques:
Dropout, Batch Normalization, Initialization Strategies for CNNs, Hyperparameter Tuning in
CNNs, Transfer Learning and Fine-tuning, Data Augmentation for CNN Training

Module 3: Advanced CNN Architectures and Applications


Object Detection with CNNs: R-CNN, Fast R-CNN, Faster R-CNN, Semantic Segmentation
with CNNs: U-Net, FCN, SegNet, Recurrent Neural Networks (RNNs) and CNN-RNN
Architectures, Generative Adversarial Networks (GANs) and CNN-GAN Architectures
Attention Mechanisms in CNNs, Image Captioning and Visual Question Answering with
CNNs

Module 4: Implementation and Deployment of CNN Models


Implementing CNNs with Deep Learning Frameworks: TensorFlow, PyTorch, Model
Deployment: Serving CNN Models for Inference, Model Optimization Techniques:
Quantization, Pruning, Model Interpretability and Visualization, Integration of CNN Models
with Web and Mobile Applications, Ethical Considerations in the Deployment of CNN
Models

Module 5: Model Evaluation and Performance Analysis


Evaluation Metrics for CNN Models: Accuracy, Precision, Recall, Cross-Validation and
Model Selection, Interpretability of CNN Predictions, Performance Optimization Strategies
for CNN Models, Case Studies and Applications of CNNs in Industry, Future Directions and
Emerging Trends in CNN Research

E. TEXTBOOKS:

"Deep Learning" by Ian Goodfellow, Yoshua Bengio, and Aaron Courville


"Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow" by Aurélien Géron
"Convolutional Neural Networks in Python: Master Data Science and Machine Learning with
Modern Deep Learning in Python, Theano, and TensorFlow" by Jason Brownlee
G. Course Articulation Matrix: (Mapping of COs with POs)

COURSE OUTCOME CORRELATION WITH PROGRAM OUTCOMES CORRELATION WITH


CO PROGRAM SPECIFIC
STATEMENT OUTCOMES
PO1 PO2 PO3 PO4 PO5 PO6 PO7 PO8 PO9 PO10 PSO1 PSO2 PSO3
CO1 Understand the
fundamental concepts
and principles of 1 1 2 1 - - 2 - 1 2 - -
Convolutional Neural
Networks (CNNs).
CO2 Learn the architecture and
components of CNNs,
including convolutional 2 3 1 2 - 2 - 1 1 1 2 1 -
layers, pooling layers, and
activation functions.
CO3 Gain proficiency in
building, training, and
evaluating CNN models 1 2 1 - 1 - - 2 3 - 1
using deep learning
frameworks.
CO4 Explore advanced topics
in CNNs, such as transfer
learning, data 1 1 2 1 - 1 - 1 1 - 2 - -
augmentation, and model
optimization.
CO5 Apply CNNs to solve
real-world computer
vision problems and 1 3 2 3 - - - - - - 1 1
analyze model
performance.
1- Low Correlation; 2- Moderate Correlation; 3- Substantial Correlation

You might also like