Wide Baseline Stereo Blog – Wide baseline stereo meets deep learning

This website is devoted to wide baseline stereo, one of the most important problems in computer vision.

ChatGPT and Image Matching

Are we done yet?

Experiments with LightGlue: geometry representation and initialization

Keep it simple, stupid

Image Matching Challenge 2023: The Unbearable Weight of the Bundle Adjustment

3D reconstruction is harder than two view matching

Un-SuperPoint family: who are they?

does it worth to know them?

Image Matching Challenge 2022 Recap

Look, LoFTRs and SuperGlue everywhere

Image Similarity Challenge 2021 recap

Can you detect edited copies?

Review of On the Limits of Pseudo Ground Truth in Visual Camera Re-localisation

Get better data!

WxBS: Relaunching challenging benchmark for Image Matching

And I mean really challenging

Submitting to IMC 2021 with custom matcher

Evaluating OpenCV new RANSACs

They become much better now

Image Matching Challenge 2020 Recap

SuperGlue, AdaLAM, DISK and others

Submitting to Image Matching Challenge 2021

Step by step tutorial

WxBS step by step

Or why it is as it is

Wide multiple baseline stereo in simple terms

Or how to explain it to everyone

Lessons learned and future directions

What I think is worthy to work on in wide baseline stereo

Benchmarking Image Retrieval for Visual Localization

Revisiting Brown patch dataset and benchmark

How to create useful development set

How to match images taken from really extreme viewpoints?

Until you succeed, try harder.

Patch extraction: devil in details

How to be smooth and fast

Local affine features: useful side product

Why to throw away an important information?

WxBS: Wide Multiple Baseline Stereo as a task

Problem definition

The Role of Wide Baseline Stereo in the Deep Learning World

Short history of wide baseline stereo in computer vision

Everything you (didn’t) want to know about image matching