Scribd
Upload
11 views8 pages

firstReport

Notes

Uploaded by

Girraj Jha
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
11 views8 pages

firstReport

Notes

Uploaded by

Girraj Jha
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/ 8

Department of Computer Science &

Engineering
Progress Report-I
Synopsis

On

Image Classification using CNNs: A Systematic Review and


Analysis

Submitted by: Submitted to:

Charu Bansal (0905CS221064) Mrs. Archana Tomar


Divykrati Niranjan (0905CS221079) Assistant Professor
Girraj Jha (0905CS221087) Dept. of CSE
Jatin Garg (0905CS221104) ITM Gwalior
Index

1. Abstract of the project


2. Introduction
3. Objectives/Aim of the project
4. Hardware and Software requirements
5. Role of team members
6. References
Abstract of the project

Convolutional Neural Networks (CNNs) are pivotal in modern computer vision, offering
unparalleled performance in image classification tasks. This research paper aim to review,
implement, and evaluate a CNN for classifying images from the CIFAR-10 dataset. The
research emphasizes CNN architecture, training methodologies, advantages, and
challenges, while exploring its potential for improvement through advanced techniques
such as deeper networks and data augmentation. This paper provides a comprehensive
overview of CNN architecture, layers, and training processes, followed by an end-to-end
application of CNN for image classification. We highlight its efficiency, advantages, and
limitations, and conclude with a discussion of future research directions.
Introduction

Introduced in the late 1980s by Yann LeCun, CNNs have evolved with advancements in
hardware and deep learning. CNN models like LeNet, AlexNet, VGG, and ResNet have
set new benchmarks in tasks such as object detection, segmentation, and recognition.

Convolutional Neural Networks (CNNs) are a type of deep learning model designed to
analyze structured data, such as images. Inspired by how the human visual system works,
CNNs are especially good at recognizing patterns and extracting important features from
data. They consist of several layers, each serving a different purpose:

1. Convolutional Layers: These layers apply filters to the input data to find specific
features like edges or textures. The filters slide over the data, creating a feature map
that highlights the presence of these features.

2. Pooling Layers: These layers reduce the size of the feature maps, making the model
faster and less likely to overfit. Pooling operations, like max pooling or average
pooling, help summarize the features in a region.

3. Fully Connected Layers: In these layers, every neuron is connected to every neuron
in the previous layer. This allows the model to combine the features extracted by
earlier layers to make predictions.

CNNs are powerful because they can automatically learn and adapt to the features of the
data, making them especially useful for tasks like image classification, object detection,
and facial recognition. They are widely used in many areas due to their strength in
handling complex data with little need for manual preprocessing.
Objectives/Aim of the project

• To understand the theoretical principles of Convolutional Neural Networks and


their architecture.

• To implement a CNN-based image classifier using the CIFAR-10 dataset.

• To optimize and evaluate the model's performance using established metrics.

• To explore challenges such as overfitting and interpretability in CNN training.

• To propose potential enhancements for better classification results.


Hardware and software requirements:
Hardware Requirements:
• Processor: Multi-core CPU (e.g., Intel i5 or better) or GPU (e.g., NVIDIA GTX 1050 or
higher).
• Memory: Minimum 8GB RAM.
• Storage: Minimum 20GB of free disk space.
• Additional: Access to cloud-based resources (e.g., Google Colab) for GPU acceleration.

Software Requirements:
• Programming Language: Python
• Libraries: TensorFlow, Keras, NumPy, Matplotlib, and Pandas.
• Tools: Jupyter Notebook or an Integrated Development Environment (IDE) like
PyCharm.
• Dataset: CIFAR-10 from TensorFlow's datasets module.
Role of team members

1. Team Lead & Performance Analyst (Charu Bansal): Oversees the project workflow, ensures
deadlines are met, and manages resources & evaluates the model's accuracy and suggests
optimization techniques.
2. Research Analyst (Girraj Jha): Investigates CNN architecture, methodologies, and best practices.
3. Data Engineer & Documentation Specialist (Jatin Garg): Prepares and preprocesses the CIFAR-
10 dataset & prepares the project report, collating findings and outcomes.
4. Model Architect (Divykrati Niranjan): Designs and implements the CNN structure using Python
and TensorFlow.
References

1. LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-Based Learning Applied to
Document Recognition. Proceedings of the IEEE, 86(11), 2278-2324.

2. Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet Classification with Deep
Convolutional Neural Networks. Advances in Neural Information Processing Systems.
3. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep Residual Learning for Image Recognition.
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.
4. R. Chauhan, K. K. Ghanshala and R. C. Joshi, "Convolutional Neural Network (CNN)
for Image Detection and Recognition," 2018 First International Conference on Secure
Cyber Computing and Communication (ICSCCC), Jalandhar, India, 2018
5. The CIFAR-10 dataset, [online] Available: https://www.cs.toronto.edu/∼kriz/cifar.html.

You might also like