# tjiiv-cprg/EPro-PnP

[CVPR 2022 Best Student Paper] EPro-PnP: Generalized End-to-End Probabilistic Perspective-n-Points for Monocular Object Pose Estimation

Repository: https://github.com/tjiiv-cprg/EPro-PnP
Canonical: https://ross.abutalabs.com/products/epro-pnp
Homepage: https://www.youtube.com/watch?v=TonBodQ6EUU
Language: Python
License: Apache-2.0
License Family: permissive
Topics: pose-estimation, 6dof, 3d-object-detection, perspective-n-point, pytorch, cvpr, monocular, levenberg-marquardt, gauss-newton
Last push: 2025-05-01T16:51:58+00:00

## Health v2 (maintenance only)
Score: 41/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 19, release rhythm 35, longevity 100
- inputs: {"age_days": 1627, "days_push": 489, "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 1175, forks 110 (observed 2026-08-28T04:03:52.462748+00:00)

## What it is
EPro-PnP is a probabilistic Perspective-n-Points (PnP) layer for end-to-end 6DoF monocular object pose estimation networks, built on PyTorch. It treats PnP as a differentiable probabilistic optimization layer, generalizable to other nested argmin learning models.

## Use cases
- estimate 6DoF object pose from a single image
- train end-to-end pose estimation networks with a differentiable PnP layer
- solve perspective-n-point with probabilistic weighting of 2D-3D correspondences
- perform monocular 3D object detection
- backpropagate through geometric optimization layers in PyTorch

## When to choose
- you need differentiable PnP inside a deep learning pipeline
- you are doing monocular 6DoF pose estimation or 3D object detection research
- you want a probabilistic alternative to categorical softmax for argmin optimization layers

## When to avoid
- you need a fast non-differentiable PnP solver for classical pipelines
- you lack a GPU or PyTorch environment
- you need production-ready real-time inference rather than research code

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, deep-learning, machine-learning, autonomous-vehicles
- platform: python
- tags: pose-estimation, 6dof, perspective-n-point, pnp, monocular, 3d-object-detection, pytorch, differentiable-optimization, cvpr-2022, levenberg-marquardt, linux, gpu

## Member repositories
- tjiiv-cprg/EPro-PnP (main) score 41

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:52.462748+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 2026-08-30T06:28:14.985277+00:00, confidence not recorded.
  - readme: https://github.com/tjiiv-cprg/EPro-PnP (fetched 2026-08-28T04:03:52.462748+00:00, sha ce412da657de)
  - homepage: https://www.youtube.com/watch?v=TonBodQ6EUU (fetched 2026-08-29T12:33:31.612911+00:00, sha 44136fa355b3)
- Data as of 2026-08-30T08:39:29.467469+00:00.