# rmislam/PythonSIFT

A clean and concise Python implementation of SIFT (Scale-Invariant Feature Transform)

Repository: https://github.com/rmislam/PythonSIFT
Canonical: https://ross.abutalabs.com/products/pythonsift
Language: Python
License: MIT
License Family: permissive
Topics: sift, template-matching, python, computer-vision, image-processing, feature-matching, opencv
Last push: 2025-08-13T05:44:52+00:00

## Health v2 (maintenance only)
Score: 48/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 36, release rhythm 35, longevity 100
- inputs: {"age_days": 4081, "days_push": 385, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1019, forks 261 (observed 2026-08-28T04:03:15.143443+00:00)

## What it is
A pure Python/NumPy implementation of SIFT (Scale-Invariant Feature Transform) that returns OpenCV KeyPoint objects and descriptors, making it a drop-in replacement for OpenCV's SIFT. It is written for clarity rather than speed and is primarily intended as an educational resource for learning how SIFT works.

## Use cases
- understand how the SIFT algorithm works internally
- extract SIFT keypoints and descriptors from images in Python
- replace OpenCV's SIFT with a readable Python implementation
- perform SIFT-based template matching
- learn computer vision feature detection through a step-by-step tutorial

## When to choose
- you want readable, educational SIFT code rather than maximum performance
- you need OpenCV-compatible keypoints and descriptors from a pure Python implementation
- you are teaching or learning the internals of scale-invariant feature detection

## When to avoid
- you need fast SIFT extraction in production - use OpenCV's optimized C++ implementation instead
- you need real-time or large-batch image processing
- you confuse it with the unrelated PySift package on PyPI

## Facets
- artifact type: library
- maturity: maintenance
- function: computer-vision, image-processing
- domain: computer-vision, image-processing, education
- platform: python
- tags: sift, feature-detection, keypoints, opencv, numpy, educational

## Member repositories
- rmislam/PythonSIFT (main) score 48

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:15.143443+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:09:24.610485+00:00, confidence not recorded.
  - readme: https://github.com/rmislam/PythonSIFT (fetched 2026-08-28T04:03:15.143443+00:00, sha 0a83fc337e8b)
- Data as of 2026-08-30T08:39:29.467469+00:00.