Current Semester Projects
1. Image Classification of All Animals using CNN
A CNN-based image classification project that identifies whether the given image belongs to any category of the animals. The model includes layers for feature extraction, pooling, and dense classification, achieving efficient binary classification results.
2. Face Recognition System Based on CNN for Criminal Identification
This project uses Convolutional Neural Networks (CNN) to identify and verify criminal faces through facial image datasets. The model is integrated into a Django web interface for real-time face recognition and identity verification.
Model Architecture (Main Model Used)
- Implemented a Convolutional Neural Network (CNN) integrated with a Django web app.
- 2D Convolutional layers + MaxPooling for progressive feature extraction.
- ReLU activation in all intermediate layers.
- Dense layers with Batch Normalization & Dropout to prevent overfitting.
- Sigmoid activation in the output layer for binary classification.
- Trained with Binary Cross-Entropy loss & evaluated via accuracy and confusion matrix.
Future Scope
In future updates, the system will be integrated with Amazon Web Services (AWS) to achieve better scalability, speed, and security. Cloud-based deployment will enable faster computation, virtualization, and cost-effective execution, making the project suitable for real-world applications.