https://github.com/namansnghl/cifar-100

Multiclass Image Classification
https://github.com/namansnghl/cifar-100

cnn-model image-processing svm-classifier

Last synced: about 1 month ago
JSON representation

Multiclass Image Classification

• Host: GitHub
• URL: https://github.com/namansnghl/cifar-100
• Owner: namansnghl
• License: mit
• Created: 2024-04-20T05:33:03.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-04-26T03:54:50.000Z (8 months ago)
• Last Synced: 2024-04-27T04:40:33.309Z (8 months ago)
• Topics: cnn-model, image-processing, svm-classifier
• Language: Jupyter Notebook
• Homepage:
• Size: 3.22 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Image Classification with CIFAR100

## About Dataset

The CIFAR-100 dataset is a collection of 60,000 32x32 colored images divided into 100 fine-grained classes. It's a popular benchmark dataset for image classification tasks, particularly in computer vision research. Here's a breakdown of its key characteristics:

- **Number of Classes**: 100 (fine-grained categories)

- **Number of Images**: 60,000 (50,000 training images, 10,000 testing images)

- **Image Size**: 32x32 pixels (RGB color channels)

- **Class Hierarchy**: The 100 classes are further organized into 20 superclasses (e.g., "mammals", "vehicles", "musical instruments"). Each image has a "fine" label (specific class) and a "coarse" label (superclass).

The fine-grained nature of the classes (e.g., distinguishing between different types of airplanes or dogs) makes CIFAR-100 a more challenging dataset compared to CIFAR-10, which has only 10 object categories.

## Methods

### Support Vector Machine

### Feed forward Neural Network

### Convolution Neural Network

## Analysis

## Results