Dog Breed Identifier

Overview

The Dog Breed Identifier is a Python program that utilizes a pre-trained ResNet50V2 model to classify images of dogs into their respective breeds. The model is fine-tuned and trained on a dataset of dog images labeled with their breeds, which are stored in the train folder and described in the labels.csv file.

How It Works

• Data Preparation:

  • The program reads breed labels from labels.csv.
  • Images from the train folder are resized and preprocessed.
  • The dataset is split into training and testing sets.

• Data Augmentation:

  • Image data augmentation is applied to the training set to enhance the model's performance.

• Model Training:

  • A ResNet50V2 model pre-trained on ImageNet is used.
  • The top layers of the model are customized for dog breed classification.
  • Training occurs over 20 epochs with callbacks for reducing learning rate and early stopping.

• Prediction:

  • An image of a dog (e.g., 'germanshepherd.jpg') can be classified into its breed using this trained model.

Files Included

Below are the files included in this project:

• dog-breed-identification.py: The main Python script that contains the code for data preparation, model training, and prediction.
• labels.csv: A CSV file that contains the mapping of image IDs to breed names.
• model: A folder that contains the saved model file (model) and the weights file (model.weights).
• test: A folder that contains three test images of different dog breeds.
• train: A folder that contains the training images of 60 dog breeds.

Running the Program

Ensure you have all required libraries installed. The following Python libraries are used in this project. Make sure to use the latest version of each for compatibility:

• TensorFlow (2.2 or higher)
• Keras (if TensorFlow version is below 2.2, install Keras separately)
• OpenCV
• Scikit-learn
• NumPy
• Pandas
• Matplotlib

You can run the program using:

python dog-breed-identification.py

This will train the model and predict the breed of a specified dog image. You can change the image path in the code to test different images. The output will print the predicted breed name for the image. For example:

Predicted Breed for this Dog is : ['leonberg']