TextBPN

Adaptive Boundary Proposal Network for Arbitrary Shape Text Detection； Accepted by ICCV2021.

Note: The complete code (including training and testing) will be released in TextBPN V2. Relevant work is advancing, and those who are interested in our work can pay more attention to the updates here.

1.Prerequisites t

python 3.9;
PyTorch 1.7.0;
Numpy >=1.2.0
CUDA 11.1;
GCC >=10.0;
NVIDIA GPU(with 11G or larger GPU memory for inference);

2.Dataset Links

3.Models

Total-Text model (pretrained on ICDAR2017-MLT)
CTW-1500 model (pretrained on ICDAR2017-MLT)
MSRA-TD500 model (pretrained on ICDAR2017-MLT)

4.Running Evaluation

run:

sh eval.sh

The details are as follows:

#!/bin/bash
##################### Total-Text ###################################
# test_size=[640,1024]--cfglib/option
CUDA_LAUNCH_BLOCKING=1 python eval_textBPN.py --exp_name Totaltext --checkepoch 390 --dis_threshold 0.3 --cls_threshold 0.825 --test_size 640 1024 --gpu 1

###################### CTW-1500 ####################################
# test_size=[640,1024]--cfglib/option
# CUDA_LAUNCH_BLOCKING=1 python eval_textBPN.py --exp_name Ctw1500 --checkepoch 560 --dis_threshold 0.3 --cls_threshold 0.8 --test_size 640 1024 --gpu 1

#################### MSRA-TD500 ######################################
# test_size=[640,1024]--cfglib/option
#CUDA_LAUNCH_BLOCKING=1 python eval_textBPN.py --exp_name TD500 --checkepoch 680 --dis_threshold 0.3 --cls_threshold 0.925 --test_size 640 1024 --gpu 1

TextBPN Adaptive Boundary Proposal Network for Arbitrary Shape Text Detection

Related tags

Overview

TextBPN

1.Prerequisites t

2.Dataset Links

3.Models

4.Running Evaluation

5.Experiments results

Owner

S.X.Zhang

AdaNet is a lightweight TensorFlow-based framework for automatically learning high-quality models with minimal expert intervention

Setup and customize deep learning environment in seconds.

Crab is a ﬂexible, fast recommender engine for Python that integrates classic information ﬁltering recommendation algorithms in the world of scientiﬁc Python packages (numpy, scipy, matplotlib).

Doing fast searching of nearest neighbors in high dimensional spaces is an increasingly important problem

HINet: Half Instance Normalization Network for Image Restoration

nn_builder lets you build neural networks with less boilerplate code

RLHive: a framework designed to facilitate research in reinforcement learning.

Implementation of Axial attention - attending to multi-dimensional data efficiently

Example of a Quantum LSTM

imbalanced-DL: Deep Imbalanced Learning in Python

Learning Skeletal Articulations with Neural Blend Shapes

PyTorch version of the paper 'Enhanced Deep Residual Networks for Single Image Super-Resolution' (CVPRW 2017)

arxiv-sanity, but very lite, simply providing the core value proposition of the ability to tag arxiv papers of interest and have the program recommend similar papers.

Models, datasets and tools for Facial keypoints detection

Official Pytorch implementation for video neural representation (NeRV)

Wenet STT Python

Code and data for paper "Deep Photo Style Transfer"

In-Place Activated BatchNorm for Memory-Optimized Training of DNNs

Framework that uses artificial intelligence applied to mathematical models to make predictions

Scalable, event-driven, deep-learning-friendly backtesting library