Rotation invariants of two dimensional curves based on iterated integrals
This provides a new method for invariant feature extraction in machine learning, specifically for applications like character recognition, though it appears incremental as it builds on existing iterated integral concepts.
The paper tackles the problem of extracting rotation-invariant features from 2D curves by introducing a novel class of invariants based on iterated integrals, presenting an algorithm for computation up to order six and applying it to online character recognition.
We introduce a novel class of rotation invariants of two dimensional curves based on iterated integrals. The invariants we present are in some sense complete and we describe an algorithm to calculate them, giving explicit computations up to order six. We present an application to online (stroke-trajectory based) character recognition. This seems to be the first time in the literature that the use of iterated integrals of a curve is proposed for (invariant) feature extraction in machine learning applications.