Skip to content

A CNN trained to distinguish between images of dogs and cats.

Notifications You must be signed in to change notification settings

SiadeBreak/AI-CNN-Dogs-or-Cats

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Is that a Dog or a Cat? 🐶😼

This "Dog or Cat" classifier is a project that applies a Convolutional Neural Network (CNN) to identify whether a photo is of a dog or a cat. The model was trained with a simple dataset of dog and cat images and achieved a 90% accuracy in classifying the images.

📊 Dataset Details

To facilitate the processing and reduce training time, this project used a reduced dataset with only 300 images of dogs and 300 images of cats. For more precise results, it is recommended to use a larger dataset. The full dataset, available in the repository, can be used to train the model with more images, increasing the accuracy and the range of breeds classified.

⚙️ Model Customization

The neural network layers can be modified to try to achieve a higher accuracy rate. Adding or removing convolutional layers, adjusting the number of filters, kernel sizes, and experimenting with different activation functions can help improve the model's performance. In this project, the current layer configuration provided the highest accuracy achieved so far.

🛠️ Technologies Used

  • Programming Language: Python

  • Libraries:

    • 🧮 NumPy
    • 👁️ OpenCV
    • 📊 Matplotlib
    • 📦 Pickle
  • Development Tools:

    • 🗒️ Google Colab
    • ☁️ Google Drive

📚 References




É um cachorro ou um gato ? 🐶😼

Esse classificador de Cachorro ou Gato é um projeto de aplicação de uma Rede Neural Convolucional (CNN) que identifica se uma foto é de um cachorro ou de um gato. O modelo foi treinado com um simples conjunto de dados de imagens de cachorros e gatos e alcançou uma precisão de 90% na classificação das imagens.

📊 Detalhes do Conjunto de Dados

Para facilitar o processamento e reduzir o tempo de treinamento, este projeto utilizou um conjunto de dados reduzido com apenas 300 imagens de cachorros e 300 imagens de gatos. Para obter resultados mais precisos, é recomendado utilizar um conjunto de dados maior. O dataset completo, disponível no repositório, pode ser utilizado para treinar o modelo com um número maior de imagens, aumentando assim a precisão e a abrangência na classificação de diferentes raças.

⚙️ Customização do Modelo

As camadas da rede neural podem ser alteradas para tentar obter uma maior taxa de acerto. Adicionar ou remover camadas convolucionais, ajustar o número de filtros, tamanhos de kernel e experimentar diferentes funções de ativação podem ajudar a melhorar a performance do modelo. Nesse projeto, a configuração atual das camadas proporcionou a maior taxa de acerto alcançada até o momento.

🛠️ Tecnologias Utilizadas

  • Linguagem de Programação: Python

  • Bibliotecas:

    • 🧮 NumPy
    • 👁️ OpenCV
    • 📊 Matplotlib
    • 📦 Pickle
  • Ferramentas de Desenvolvimento:

    • 🗒️ Google Colab
    • ☁️ Google Drive

📚 Referências

About

A CNN trained to distinguish between images of dogs and cats.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published