WenmuZhou/DBNet.pytorch
A PyTorch implementation of DBNet, a real-time scene text detection model using differentiable binarization.

Velocity · 7d
+0.4
★ / day
Trend
→steady
star history
This repository provides a PyTorch re-implementation of the DBNet model for real-time scene text detection. The model uses a ResNet backbone with Feature Pyramid Network (FPN) and a differentiable binarization head to detect text regions in natural scene images. It supports single and multi-GPU training with standard dataset formats for training and evaluation on ICDAR-style benchmarks.