🤗 image matching toolbox webui

This repository contains ROS2 package for 2d-Lidar scanner based Kidnapped Robot Re-localization solution (without using AMCL). The idea is to take the original map image, and create a scan-image from local lidar scan, then 'simulate' scans on random positions in the map image (using ray-casting), and find the best-matching scan with orig scan.

[CVPR'24] MESA: Matching Everything by Segmenting Anything

Visual localization made easy with hloc

[CVPR 2024] RoMa: Robust Dense Feature Matching; RoMa is the robust dense feature matcher capable of estimating pixel-dense warps and reliable certainties for almost any image pair.

xLPM: xGausaain-Splatting Models with Localized Point Management

ONNX-compatible LightGlue: Local Feature Matching at Light Speed. Supports TensorRT, OpenVINO

This repository provides a collection of Python tools and applications for comparing images using various algorithms such as ORB (Oriented FAST and Rotated BRIEF) and SSIM (Structural Similarity Index). The tools include both command-line scripts and graphical user interface (GUI) applications for ease of use

[WACV 2024] BALF: Simple and Efficient Blur Aware Local Feature Detector

SuperGlue: Learning Feature Matching with Graph Neural Networks (CVPR 2020, Oral)

This project provides a tool to compare two images using various similarity metrics, including histograms, structural similarity index (SSIM), mean squared error (MSE), mean absolute error (MAE), feature matching, and image hashing.

image stitching using superglue

Panorama image stitching using classical and Deep learning computer vision methods

Make Rendered Database using Rendered Image Obtained by Trained Gaussian Splatting

3D scene reconstruction and camera pose estimation from custom dataset images

API for panoramer

Stabilo is a Python package for stabilizing video frames or object trajectories using advanced transformation techniques. It supports user-defined masks to exclude specific areas, making it ideal for dynamic scenes with moving objects. Unlike traditional stabilization, it stabilizes content relative to a chosen reference frame.

Code for "LGFCTR: Local and Global Feature Transformer for Image Matching"

Patch2Pix: Epipolar-Guided Pixel-Level Correspondences [CVPR2021]

[AAAI2023] TopicFM: Robust, Efficient, and Interpretable Topic-Assisted Feature Matching