https://github.com/beenish-ishtiaq/dep-task-4-image-classification-cifar10

Developed a Convolutional Neural Network (CNN) model to classify images into 10 categories. The project includes data augmentation, model building, training, and evaluation.
https://github.com/beenish-ishtiaq/dep-task-4-image-classification-cifar10

cifar10 convolutional-neural-networks data-augmentation deep-learning dropout image-classification machine-learning model-training python regularization

Last synced: 8 months ago
JSON representation

Developed a Convolutional Neural Network (CNN) model to classify images into 10 categories. The project includes data augmentation, model building, training, and evaluation.

• Host: GitHub
• URL: https://github.com/beenish-ishtiaq/dep-task-4-image-classification-cifar10
• Owner: beenish-Ishtiaq
• Created: 2024-07-28T13:35:43.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2024-07-28T13:44:36.000Z (about 1 year ago)
• Last Synced: 2025-01-11T13:28:46.011Z (9 months ago)
• Topics: cifar10, convolutional-neural-networks, data-augmentation, deep-learning, dropout, image-classification, machine-learning, model-training, python, regularization
• Language: Jupyter Notebook
• Homepage: https://www.linkedin.com/posts/beenish-ishtiaq-a4652a237_machinelearning-deeplearning-imageclassification-activity-7223326306724171776-DrFD?utm_source=share&utm_medium=member_desktop
• Size: 169 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Image Classification using CIFAR-10

## Project Overview

In this project, a Convolutional Neural Network (CNN) is being developed to classify images from the CIFAR-10 dataset into 10 predefined categories. The main steps involved data preprocessing, data augmentation, model building, training, and evaluation.

## Dataset

The CIFAR-10 dataset consists of 60,000 32x32 color images in 10 classes, with 6,000 images per class. There are 50,000 training images and 10,000 test images. The dataset includes the following classes:

- Airplane

- Automobile

- Bird

- Cat

- Deer

- Dog

- Frog

- Horse

- Ship

- Truck

## Key Steps

1. **Data Augmentation and Preprocessing**: 

    - Applied techniques such as horizontal flip, random rotation, and random zoom to augment the training data and prevent overfitting.

    - Normalized pixel values to be between 0 and 1.

2. **Building the Convolutional Neural Network (CNN)**: 

    - Constructed a CNN model with multiple convolutional layers, max pooling layers, dropout layers, and dense layers.

3. **Training the Model**: 

    - Trained the model using the training dataset with data augmentation.

4. **Model Evaluation and Fine-tuning**: 

    - Evaluated the model using the test dataset and plotted the training/validation accuracy and loss over epochs.

5. **Handling Overfitting**: 

    - Used dropout and regularization techniques to mitigate overfitting.

## Model Architecture

The CNN model used in this project has the following architecture:

- Input layer: 32x32x3 (CIFAR-10 images)

- Data Augmentation layer

- Conv2D layer: 32 filters, (3,3) kernel, ReLU activation

- MaxPooling2D layer: (2,2) pool size

- Conv2D layer: 64 filters, (3,3) kernel, ReLU activation

- MaxPooling2D layer: (2,2) pool size

- Conv2D layer: 128 filters, (3,3) kernel, ReLU activation

- MaxPooling2D layer: (2,2) pool size

- Flatten layer

- Dense layer: 512 units, ReLU activation

- Dropout layer: 0.5 dropout rate

- Dense layer: 10 units, softmax activation

## Results

The model achieved an accuracy of 68% on the test dataset. Further tuning and experimentation can be done to improve the model performance.