When incoming students come to the UIUC campus for the first time and attend classes in the first few weeks, they usually have a difficult time recognizing buildings and have very little knowledge about them. Our web-app WhereAmI helps students identify buildings and provides them with useful information about the buildings in the engineering campus.
graph TD
A[Input Image] --> B[Pretained ResNet50 Model]
B --> C[Fine-tuning Layers]
C --> D[Customize last Fully Connected Layer]
D --> E[Building Classification]
style A fill:#f9f,stroke:#333,stroke-width:2px
style E fill:#bbf,stroke:#333,stroke-width:2px
- Random Crop: 256px, scale=(0.8, 1.0)
- Rotation: ±30 degrees
- Horizontal Flip: p=0.5
- Color Jitter: brightness=0.2, contrast=0.2
- Grayscale: p=0.1
- Modified final fully connected layer
- BatchNormalization
- Dropout (0.4) for regularization
- Fine-tuned last two residual blocks
- Batch Size: 16
- Learning Rate: 0.0005
- Optimizer: AdamW
- Weight Decay: 0.05
- Early Stopping: 5 epochs
- Training Accuracy: above 95%
- Validation Accuracy: near 80%
- Epoches: 20
Fiona Bai, Shaoshao Xiong
- Primary Role: Working on image data processing and frontend development.
- Skills Focus: JavaScript, CSS, HTML, Django.
Evelyn Zhou, Jay Chen
- Primary Role: Training the CNN model and managing the database using Django.
- Skills Focus: Deep learning, Pytorch, Django.
- Collecting images.
- Using Django to connect the front and back end code.
- Flexible roles with task exchanges likely as the project progresses.
- Ensure you have Node.js (version 16.x or above recommended) installed.
- Ensure you have npm (Node Package Manager) installed.
- Install Node.js Dependencies
Navigate to the project root directory and run the following command:npm install
In the root directory (where manage.py is located), run the Django development server:
python manage.py runserverThe app should now be accessible at:
http://127.0.0.1:8000