Non Linear Leastsquares solver:
- Implemented as a templated class in a single header
- May be used to determine any number of params (given enough data points)
- Enable use of robust kernels for outlier rejection
- Allows constraints to curve fitting params
- Uses paralelism with OMP for faster convergence
Compiled in windows from Visual-Studio-code + cmake with MSVC and msys2. Uses Eigen header only library, included in project.
- main.cpp file contains sample data table (y,x) of a noisy gaussian distribution
- The solver is used to determine the best p's for fitting data to function: y = f(x) = p0 * exp(-pow((x - p1) * p2, 2))