AnimalVision is a deep learning project that uses a Convolutional Neural Network (CNN) to classify animal images. With a dataset of thousands of images and 90 different animal classes, the model achieves high accuracy in predictions.
- Model:
AnimalVisionModel.keras - Dataset: 5,400 images across 90 animal classes
- Architecture: CNN-based classification model
- Performance Metrics:
- Validation Accuracy: 85%
- AUC Score: 99%
This model has been trained to recognize a diverse set of animals with high precision.
You can test the model using our Streamlit web demo:
🔗 AnimalVision Demo
- Upload an image of an animal.
- The model will analyze the image and return the top predicted class.
- A bar chart will display the top 5 predictions with their confidence scores.
If you want to run the project locally, follow these steps:
git clone https://github.com/furkankarakuz/AnimalVision.git
cd AnimalVisionMake sure you have Python installed, then install the required packages:
pip install -r requirements.txtstreamlit run app.pyNow you can access the app locally and test images directly from your browser!
The AnimalVisionModel was trained on a dataset of 5,400 images across 90 animal categories. The training process involved:
- Data augmentation techniques to improve generalization.
- A CNN architecture optimized for image classification.
- Evaluation using accuracy and AUC metrics.
With a validation accuracy of 85% and an AUC score of 99%, the model demonstrates strong classification capabilities.
Below is an example output when testing an image:
| Rank | Animal | Confidence |
|---|---|---|
| 1️⃣ | Tiger 🐯 | 92% |
| 2️⃣ | Lion 🦁 | 87% |
| 3️⃣ | Cheetah 🐆 | 79% |
| 4️⃣ | Leopard 🐆 | 73% |
| 5️⃣ | Jaguar 🐆 | 69% |
This example demonstrates how the model ranks the possible classifications with confidence scores.
Contributions are welcome! If you have ideas for improvement or want to enhance the model, feel free to fork the repository and submit a pull request.
This project is licensed under the Apache-2.0 License. See the LICENSE file for more details.
🐾 Explore the world of animals with AI! 🎉