Inlier

inlier is a rust library for robust model fitting. It exposes idiomatic Rust API with pluggable estimators, samplers, scoring functions, and optimizers.

Features

• Drop-in estimators for homography, fundamental/essential matrices, absolute pose, rigid transform, etc.
• Ready-to-use RANSAC variants (RANSAC, MSAC, MAGSAC, ACRANSAC) with local optimization pipelines (LSQ, IRLS, nested RANSAC, etc.).
• Flexible samplers (uniform, PROSAC, NAPSAC, progressive NAPSAC, adaptive reordering) plus neighborhood graphs (grid, USearch ANN).

Run Examples

# Robust line fitting with plot output
cargo run --example linear_regression

# Homography estimation demo
cargo run --example homography_estimation

See examples/README.md for the full catalog.

Usage

use inlier::{estimate_homography, RansacSettings};
use nalgebra::DMatrix;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let points1 = DMatrix::from_row_slice(4, 2, &[0.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0]);
    let points2 = DMatrix::from_row_slice(4, 2, &[1.0, 1.0, 2.0, 1.0, 2.0, 2.0, 1.0, 2.0]);

    let result = estimate_homography(&points1, &points2, 1.0, Some(RansacSettings::default()))?;
    println!("Inliers: {}", result.inliers.len());
    Ok(())
}

Extensibility

Implement the core traits to add custom functionality:

• Estimator (models)
• Sampler / NeighborhoodGraph
• Scoring
• LocalOptimizer
• TerminationCriterion

Each trait includes documentation and doctests describing the expected behavior.

Reference

This uses the c++ library superansac as reference implementation, of which the author had done an excellent work on improving the SOTA sample consensus algorithm!