Classificacao_de_Gatos_e_Cachorros_com_Transfer_Learning_VGG16

Classification of Cats and Dogs with Transfer Learning (VGG16) This project utilizes the Transfer Learning technique with the pre-trained VGG16 model to classify images of cats and dogs.

Code Structure

• Importing Libraries: Import necessary libraries such as TensorFlow, Keras, NumPy, and Matplotlib.
• Dataset Preparation: The image dataset is organized into specific directories (cats and dogs).
• Image Processing: Random selection of up to 50 images per class, resizing, and normalization.
• Data Shuffling and Splitting: The data is shuffled and split into training (70%), validation (15%), and test (15%) sets.
• Pre-trained VGG16 Model: Loading the VGG16 model with ImageNet weights and adding custom layers.
• Training and Evaluation: Training the model and evaluating its performance on the test set.
• Visualization of Results: Plotting the training and validation accuracy over epochs.

Results

Test Set Accuracy: Varies based on training The model achieved an accuracy of approximately 73% on the test set in one execution, demonstrating a good ability to classify images of cats and dogs. The accuracy may vary depending on dataset characteristics and training conditions.

Requirements:

• TensorFlow
• Keras
• NumPy
• Matplotlib

Visualization of Results

The training and validation accuracy graphs over epochs were plotted for visual analysis of the model's performance.

Main Points of the Code

• Image Processing: Selection of up to 50 random images of cats and dogs, resizing, and normalization.
• Data Splitting: The data was split into training (70%), validation (15%), and test (15%) sets.
• Pre-trained VGG16 Model: Utilization of the VGG16 model with ImageNet weights and addition of custom layers for classification.
• Training: The model was trained for 10 epochs, monitoring validation accuracy.
• Evaluation: Achieved approximately 73% accuracy on the test set in one execution, but the result may vary.

Result Summary

After training the image classification model using transfer learning with VGG16, the performance was evaluated on the test set. The model achieved an accuracy of around 73%, but the result may vary depending on training conditions and dataset characteristics. This demonstrates a good ability of the model to differentiate between the two classes.