Path Finding via Shape Context Matching

The popularity of sports has led to a surge in sports-related apps. While numerous running apps are capable of setting goals, routes, and tracking sports activity, they lack the ability to automatically map out routes tailored to user preferences, such as routes with specific or interesting shapes.Given the dearth of previous studies on comparing paths based on their shape or altitude, this paper proposes an algorithm that can identify a route satisfying user specifications within a specified region. The paper comprises one major topic, shape matching path. Dealing with the 2D scenario, the initial step is to extract the feature points of a desired shape specified by the user and a background image, and perform feature point matching using the shape context after removing any noise. This yields a closed route in a given area that approximates the outline of the desired shape.