As an AI Researcher in a school laboratory project, I collaborated with TicketWindow.ca to develop an AI-powered solution for automating seat detection from venue seating maps. The project focused on leveraging computer vision and object detection techniques to accurately identify and count seats in each row and section of a venue.

Key Contributions & Impact

• Pre-trained Model Research and Selection: Researched and evaluated multiple pre-trained object detection models to determine the most accurate and efficient model for the seat detection task. This involved assessing performance metrics and considering the specific characteristics of the venue map images.
• AI-Powered Seat Detection and Counting: Developed an AI solution to automatically detect and count seats within each row and section of venue images.

Challenges & Solutions

• Variations in Venue Map Styles: Venue maps can vary significantly in terms of style, layout, color schemes, and the way seats are represented. Solution: Trained the object detection model on a diverse dataset of venue maps representing different styles and layouts. Utilized data augmentation techniques to increase the model's robustness to variations in appearance.

Lessons Learned

• Application of AI in a Real-World Context: Gained insights into the application of AI in a real-world scenario, working with a company (TicketWindow.ca) to solve a practical problem.