Evaluation Rubric

Model Performance

50% (Max: 50 points)

• Accuracy on real test data (e.g., mAP, precision/recall, IoU)

1-30

• Performance consistency across object classes (avoiding overfitting)

1-15

• Effective use of synthetic data to generalize

1-5

Presentation of Solution

20% (Max: 20 points)

• Problem and approach were clearly defined

1-5

• Presentation was structured and easy to follow

1-5

• Visuals effectively explained challenges and results

1-5

• Delivery was confident and engaging

1-5

Model Interpretability and Insights

15% (Max: 15 points)

• Interpretation of model predictions and errors

1-5

• Identified key performance bottlenecks

1-5

• Addressed challenges of domain adaptation

1-5

Collaboration and Teamwork

15% (Max: 15 points)

• Effective division of tasks and collaboration

1-5

• Problem-solving abilities under pressure

1-5

• Adaptability to new information or constraints

1-5