Urban community planning often relies on top-down strategies that fail to account for the spontaneous ways residents actually use public spaces. This disconnect frequently leads to conflicts between original design intentions and the reality of daily life. To address this, the paper "CommuniWave:A Machine Learning Model for Quantifying the Degree of Temporary Informal Behavior in Urban Communities" introduces a new tool designed to help urban managers and designers better understand and quantify these informal activities. By measuring how residents interact with their environment, the model aims to support more refined, data-driven decision-making to improve community resilience.
Bridging the Gap in Urban Planning
Current urban design processes often lack effective metrics to track how people use spaces in ways not originally planned. This lack of data makes it difficult for managers to adapt to the complexity and uncertainty of urban life. CommuniWave seeks to solve this by providing a way to monitor "informal behaviors" dynamically, allowing planners to move away from rigid, top-down approaches and toward designs that better reflect the actual needs and habits of the community.
How the Model Works
CommuniWave functions as an integrated machine learning pipeline that processes street-level video footage to extract meaningful data. The system is built on three core components:
Behavior Capture Net (BCN): Utilizes mmaction2 to identify and capture specific actions occurring in the video.
YLX (YOLOv10): A self-developed version of the YOLOv10 model used to detect and track objects or individuals within the urban environment.
Behavior Eval Model (BEM): A random forest-based model that evaluates the captured data to determine the Degree of Informal Behavior (DIB).
Monitoring and Decision Support
The final output of the CommuniWave model is a series of DIB fluctuation charts. These charts provide a visual representation of how informal behavior changes over time within a specific community. By transforming raw street video into actionable data, the model enables urban managers to monitor their spaces continuously. This dynamic feedback loop is intended to help officials make more informed, refined decisions that enhance the overall resilience of urban environments in the face of changing community needs.
Comments (0)
to join the discussion
No comments yet
Be the first to share your thoughts!