拼接和去重

https://github.com/magicleap/SuperGluePretrainedNetwork

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

 View license

 1k stars  187 forks

 Star

 Watch

• Code
• Issues7
• Pull requests
• Actions
• Projects
• Security
• Insights

 

 master 

 

Go to file

Latest commit

zmurez-ml Merge pull request #30 from ddetone/fix_phototourism_test_pairs

…

Aug 26, 2020

Git stats

•  11

Files

View code

README.md

Research @ Magic Leap (CVPR 2020, Oral)

SuperGlue Inference and Evaluation Demo Script

Introduction

SuperGlue is a CVPR 2020 research project done at Magic Leap. The SuperGlue network is a Graph Neural Network combined with an Optimal Matching layer that is trained to perform matching on two sets of sparse image features. This repo includes PyTorch code and pretrained weights for running the SuperGlue matching network on top of SuperPointkeypoints and descriptors. Given a pair of images, you can use this repo to extract matching features across the image pair.

SuperGlue operates as a "middle-end," performing context aggregation, matching, and filtering in a single end-to-end architecture. For more details, please see:

• Full paper PDF: SuperGlue: Learning Feature Matching with Graph Neural Networks.

• Authors: Paul-Edouard Sarlin, Daniel DeTone, Tomasz Malisiewicz, Andrew Rabinovich

• Website: psarlin.com/superglue for videos, slides, recent updates, and more visualizations.

• hloc: a new toolbox for visual localization and SfM with SuperGlue, available at cvg/Hierarchical-Localization. Winner of 3 CVPR 2020 competitions on localization and image matching!

We provide two pre-trained weights files: an indoor model trained on ScanNet data, and an outdoor model trained on MegaDepth data. Both models are inside the weights directory. By default, the demo will run theindoor model.

Dependencies

• Python 3 >= 3.5
• PyTorch >= 1.1
• OpenCV >= 3.4 (4.1.2.30 recommended for best GUI keyboard interaction, see this note)
• Matplotlib >= 3.1
• NumPy >= 1.18

Simply run the following command: pip3 install numpy opencv-python torch matplotlib

Contents

There are two main top-level scripts in this repo:

1. demo_superglue.py : runs a live demo on a webcam, IP camera, image directory or movie file
2. match_pairs.py: reads image pairs from files and dumps matches to disk (also runs evaluation if ground truth relative poses are provided)

Live Matching Demo Script (demo_superglue.py)

This demo runs SuperPoint + SuperGlue feature matching on an anchor image and live image. You can update the anchor image by pressing then key. The demo can read image streams from a USB or IP camera, a directory containing images, or a video file. You can pass all of these inputs using the --input flag.

Run the demo on a live webcam

Run the demo on the default USB webcam (ID #0), running on a CUDA GPU if one is found:

./demo_superglue.py

Keyboard control:

• n: select the current frame as the anchor
• e/r: increase/decrease the keypoint confidence threshold
• d/f: increase/decrease the match filtering threshold
• k: toggle the visualization of keypoints
• q: quit

Run the demo on 320x240 images running on the CPU:

./demo_superglue.py --resize 320 240 --force_cpu

The --resize flag can be used to resize the input image in three ways:

1. --resize width height : will resize to exact width x heightdimensions
2. --resize max_dimension : will resize largest input image dimension to max_dimension
3. --resize -1 : will not resize (i.e. use original image dimensions)

The default will resize images to 640x480.

Run the demo on a directory of images

The --input flag also accepts a path to a directory. We provide a directory of sample images from a sequence. To run the demo on the directory of images in freiburg_sequence/ on a headless server (will not display to the screen) and write the output visualization images todump_demo_sequence/:

./demo_superglue.py --input assets/freiburg_sequence/ --output_dir dump_demo_sequence --resize 320 240 --no_display

You should see this output on the sample Freiburg-TUM RGBD sequence:

The matches are colored by their predicted confidence in a jet colormap (Red: more confident, Blue: less confident).

Additional useful command line parameters

• Use --image_glob to change the image file extension (default:*.png, *.jpg, *.jpeg).
• Use --skip to skip intermediate frames (default: 1).
• Use --max_length to cap the total number of frames processed (default: 1000000).
• Use --show_keypoints to visualize the detected keypoints (default:False).

Run Matching+Evaluation (match_pairs.py)

This repo also contains a script match_pairs.py that runs the matching from a list of image pairs. With this script, you can:

• Run the matcher on a set of image pairs (no ground truth needed)
• Visualize the keypoints and matches, based on their confidence
• Evaluate and visualize the match correctness, if the ground truth relative poses and intrinsics are provided
• Save the keypoints, matches, and evaluation results for further processing
• Collate evaluation results over many pairs and generate result tables

Matches only mode

The simplest usage of this script will process the image pairs listed in a given text file and dump the keypoints and matches to compressed numpy npz files. We provide the challenging ScanNet pairs from the main paper in assets/example_indoor_pairs/. Running the following will run SuperPoint + SuperGlue on each image pair, and dump the results todump_match_pairs/:

./match_pairs.py

The resulting .npz files can be read from Python as follows:

>>> import numpy as np
>>> path = 'dump_match_pairs/scene0711_00_frame-001680_scene0711_00_frame-001995_matches.npz'
>>> npz = np.load(path)
>>> npz.files
['keypoints0', 'keypoints1', 'matches', 'match_confidence']
>>> npz['keypoints0'].shape
(382, 2)
>>> npz['keypoints1'].shape
(391, 2)
>>> npz['matches'].shape
(382,)
>>> np.sum(npz['matches']>-1)
115
>>> npz['match_confidence'].shape
(382,)

For each keypoint in keypoints0, the matches array indicates the index of the matching keypoint in keypoints1, or -1 if the keypoint is unmatched.

Visualization mode

You can add the flag --viz to dump image outputs which visualize the matches:

./match_pairs.py --viz

You should see images like this inside of dump_match_pairs/ (or something very close to it, see this note):

The matches are colored by their predicted confidence in a jet colormap (Red: more confident, Blue: less confident).

Evaluation mode

You can also estimate the pose using RANSAC + Essential Matrix decomposition and evaluate it if the ground truth relative poses and intrinsics are provided in the input .txt files. Each .txt file contains three key ground truth matrices: a 3x3 intrinsics matrix of image0: K0, a 3x3 intrinsics matrix of image1: K1 , and a 4x4 matrix of the relative pose extrinsics T_0to1.

To run the evaluation on the sample set of images (by default readingassets/scannet_sample_pairs_with_gt.txt), you can run:

./match_pairs.py --eval

Since you enabled --eval, you should see collated results printed to the terminal. For the example images provided, you should get the following numbers (or something very close to it, see this note):

Evaluation Results (mean over 15 pairs):
AUC@5    AUC@10  AUC@20  Prec    MScore
26.99    48.40   64.47   73.52   19.60

The resulting .npz files in dump_match_pairs/ will now contain scalar values related to the evaluation, computed on the sample images provided. Here is what you should find in one of the generated evaluation files:

>>> import numpy as np
>>> path = 'dump_match_pairs/scene0711_00_frame-001680_scene0711_00_frame-001995_evaluation.npz'
>>> npz = np.load(path)
>>> print(npz.files)
['error_t', 'error_R', 'precision', 'matching_score', 'num_correct', 'epipolar_errors']

You can also visualize the evaluation metrics by running the following command:

./match_pairs.py --eval --viz

You should also now see additional images in dump_match_pairs/ which visualize the evaluation numbers (or something very close to it, see thisnote):

The top left corner of the image shows the pose error and number of inliers, while the lines are colored by their epipolar error computed with the ground truth relative pose (red: higher error, green: lower error).

Running on sample outdoor pairs

[Click to expand]

Recommended settings for indoor / outdoor

[Click to expand]

Test set pair file format explained

[Click to expand]

Reproducing the indoor evaluation on ScanNet

[Click to expand]

Reproducing the outdoor evaluation on YFCC

[Click to expand]

Reproducing outdoor evaluation on Phototourism

[Click to expand]

Correcting EXIF rotation data in YFCC and Phototourism

[Click to expand]

Outdoor training / validation scene splits of MegaDepth

[Click to expand]

A note on reproducibility

[Click to expand]

Creating high-quality PDF visualizations and faster visualization with --fast_viz

[Click to expand]

BibTeX Citation

If you use any ideas from the paper or code from this repo, please consider citing:

@inproceedings{sarlin20superglue,
  author    = {Paul-Edouard Sarlin and
               Daniel DeTone and
               Tomasz Malisiewicz and
               Andrew Rabinovich},
  title     = {{SuperGlue}: Learning Feature Matching with Graph Neural Networks},
  booktitle = {CVPR},
  year      = {2020},
  url       = {https://arxiv.org/abs/1911.11763}
}

Additional Notes

• For the demo, we found that the keyboard interaction works well with OpenCV 4.1.2.30, older versions were less responsive and the newest version had a OpenCV bug on Mac
• We generally do not recommend to run SuperPoint+SuperGlue below 160x120 resolution (QQVGA) and above 2000x1500
• We do not intend to release the SuperGlue training code.
• We do not intend to release the SIFT-based or homography SuperGlue models.

Legal Disclaimer

Magic Leap is proud to provide its latest samples, toolkits, and research projects on Github to foster development and gather feedback from the spatial computing community. Use of the resources within this repo is subject to (a) the license(s) included herein, or (b) if no license is included, Magic Leap's Developer Agreement, which is available on our Developer Portal. If you need more, just ask on the forums! We're thrilled to be part of a well-meaning, friendly and welcoming community of millions.

Releases

No releases published

Packages

No packages published 

Contributors 3

• zmurez-ml
• Skydes Paul-Edouard Sarlin
• hiakru WESLEY (Wei Li)

Languages

 

• Python100.0%