Team
Prachi Khatri, Parul Batish, Mohammed Bilal Khan, Maharshi Mehta
Categories
Artificial Intelligence
Overview
The Project analysis compared YOLO v8, YOLO v9, and UNet for image analysis tasks, focusing on object detection and gap classification. YOLO v8 showed superior accuracy in object detection, while YOLO v9 performed moderately in image segmentation. UNet excelled in accurately classifying gap types, achieving a remarkable 97% accuracy, surpassing both YOLO models.
The findings underscore the significance of considering various factors like dataset quality and model architecture in determining performance. Despite YOLO's strength in object detection, UNet's specialized focus on image segmentation proved highly effective for the specific task of gap classification. Further research and experimentation are encouraged to harness the potential of different models and improve overall performance.
Semantic Segmentation of Sidewalk for Gap Detection- Kevares
Maintaining pedestrian accessibility and safety in urban settings is crucial, yet wear and tear on sidewalks due to factors like foot traffic, weather, and deteriorating infrastructure poses significant challenges. Maintaining safe and accessible pathways for all community members requires prompt identification and resolution of sidewalk issues.
