Landmark Classification Project

This project is aimed at classifying images of landmarks using a deep learning model built with Keras and VGG19. The dataset contains images of various landmarks, each labeled with a unique identifier.

Table of Contents

• Installation
• Usage
• Dataset
• Model Architecture
• Training
• Evaluation
• Results
• Contributing
• License

Installation

Prerequisites

• Python 3.x
• Required Python packages:
  • numpy
  • pandas
  • keras
  • tensorflow
  • opencv-python
  • matplotlib
  • scikit-learn
  • pillow

Installing

Clone the repository and install the required packages:

git clone https://github.com/your-username/landmark-classification.git
cd landmark-classification
pip install -r requirements.txt

Usage

1. Prepare your dataset and place the train.csv file in the root directory.
2. Run the script to preprocess the data and train the model.

python train.py

Dataset

The dataset should include a train.csv file containing the image file names and their corresponding landmark IDs. Images should be organized in folders following the structure derived from their filenames.

• train.csv - CSV file with two columns: fname (image file name) and landmark_id (class label).
• Images should be in subdirectories named based on the first three characters of their filenames.

Model Architecture

The model is based on the VGG19 architecture with some modifications:

• Batch Normalization after specific layers
• Dropout layers for regularization
• Dense layer at the end for classification

Training

The training script train.py preprocesses the images, encodes the labels, and trains the model. The model is trained using the RMSprop optimizer and the sparse categorical crossentropy loss function.

Training Parameters

• Batch size: 16
• Epochs: 1 (can be adjusted)
• Learning rate: 0.0001
• Data augmentation is applied to training images.

Training Script

python train.py

Evaluation

The model is evaluated on a validation set, and the accuracy and loss are plotted for both training and validation sets. Misclassified images are stored for further analysis.

Evaluation Script

python evaluate.py

Results

The results of the training and evaluation process, including accuracy and loss plots, are displayed using matplotlib.

Contributing

Contributions are welcome! Please fork the repository and submit a pull request with your changes.

License

This project is licensed under the MIT License - see the LICENSE file for details.