📚 A study on Image Classification

This repository aims to explore fundamental concepts of image classification by implementing popular architectures throughout history. In total, we selected ten approaches to the task of image classification, to study them and how they propose to solve this classical problem of computer vision.

Each architecture is implemented in a jupyter environment with pytorch, being trained and tested on three popular datasets for the task. The specified datasets for evaluating the models are MNIST, CIFAR-100, and Imagenet. This repository is divided into folders for each selected approach, containing the technical details of the models, their implementation, and the results of their performance on the tests.

📌 Approaches selected for study

• LeNet

• AlexNet

• GoogLeNet

• VGG

• ResNet

• ResNeXt

• Inception

• ViT

• Swin

• ConvNeXt

📑 References