woctezuma / feature-matcher

Match similar image features.

🎯 Feature Matcher

This repository contains Python code to match representation vectors, using cosine similarity.

Requirements

• Install the latest version of Python 3.X.
• Install the required packages:

pip install -r requirements.txt

Usage

Run:

!python match_fts.py \
 --input_dir features \
 --feature_filename fts.npy \
 --output_dir matches \
 --numpy_matches matches.npy \
 --numpy_similarity_scores scores.npy \
 --num_neighbors 10

The following files will be saved in the --output_dir folder:

• matches.npy: the matched indices (as np.uint32) in a NumPy file,
• scores.npy: the similarity scores (as np.float16) in a NumPy file.

Example

For instance, to match features for images in the balloon dataset:

%cd /content
!git clone https://github.com/woctezuma/feature-extractor.git
%cd feature-extractor
%pip install --quiet -r requirements.txt

!wget https://github.com/matterport/Mask_RCNN/releases/download/v2.1/balloon_dataset.zip
!unzip -q balloon_dataset.zip

!python extract_fts.py --data_dir balloon

%cd /content
!git clone https://github.com/woctezuma/feature-matcher.git
%cd feature-matcher
%pip install --quiet -r requirements.txt

!python match_fts.py --input_dir /content/feature-extractor/features

References

• A feature extractor for Github repositories which include a hubconf.py file.
• match-steam-banners: retrieve games with similar store banners.
• steam-DINOv2: retrieve games with similar store banners, using Meta AI's DINOv2.