How is the lensed image computed? For each output pixel at angular position θ, this solves the lens equation β = θ − α(θ) to find the source-plane position β, then samples the source image there (inverse ray tracing). The deflection α is computed from a Navarro-Frenck-white lens mass profile.