# MhLiao/DB

A PyTorch implementation of "Real-time Scene Text Detection with Differentiable Binarization".

Repository: https://github.com/MhLiao/DB
Canonical: https://ross.abutalabs.com/products/mhliao-db
Language: Python
License Family: other
Topics: scene-text-detection, dbnet
Last push: 2024-03-11T06:19:46+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2480, "days_push": 905, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2260, forks 486 (observed 2026-08-28T04:06:31.738848+00:00)

## What it is
A PyTorch implementation of DBNet and DBNet++, real-time arbitrary-shape scene text detection models based on differentiable binarization. It includes training, evaluation, demo scripts, and pretrained models.

## Use cases
- detect text in natural scene images
- train a scene text detection model
- run real-time text detection on images
- extract text regions from photos before OCR
- reproduce DBNet research results
- use pretrained DBNet++ models

## When to choose
- you need arbitrary-shape scene text detection in PyTorch
- you want a research-grade DBNet implementation with pretrained weights
- you need real-time text detection performance

## When to avoid
- you need a production OCR pipeline rather than a research codebase
- you want a maintained library with a license and modern PyTorch support
- you need text recognition, not just detection

## Facets
- artifact type: library
- maturity: maintenance
- function: computer-vision, ocr, machine-learning, deep-learning
- domain: computer-vision, image-processing, machine-learning
- platform: python
- tags: dbnet, dbnet-plus, scene-text-detection, pytorch, text-detection, research-code, natural-language-processing, linux, gpu

## Member repositories
- MhLiao/DB (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:31.738848+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 2026-08-30T02:42:38.525529+00:00, confidence not recorded.
  - readme: https://github.com/MhLiao/DB (fetched 2026-08-28T04:06:31.738848+00:00, sha 2c9b563dea06)
- Data as of 2026-08-30T08:39:29.467469+00:00.