STEFANN: Scene Text Editor using Font Adaptive Neural Network

Overview

Getting Started  •   Training Networks  •   External Links  •   Citation  •   License



The official GitHub repository for the paper on STEFANN: Scene Text Editor using Font Adaptive Neural Network.


Getting Started

1. Installing Dependencies

Package Source Version Tested version
(Updated on April 14, 2020)
Python Conda 3.7.7 ✔️
Pip Conda 20.0.2 ✔️
Numpy Conda 1.18.1 ✔️
Requests Conda 2.23.0 ✔️
TensorFlow Conda 2.1.0 ✔️
Keras Conda 2.3.1 ✔️
Pillow Conda 7.0.0 ✔️
Colorama Conda 0.4.3 ✔️
OpenCV PyPI 4.2.0 ✔️
PyQt5 PyPI 5.14.2 ✔️

💥 Quick installation

Step 1: Install Git and Conda package manager (Miniconda / Anaconda)

Step 2: Update and configure Conda

conda update conda
conda config --set env_prompt "({name}) "

Step 3: Clone this repository and change directory to repository root

git clone https://github.com/prasunroy/stefann.git
cd stefann

Step 4: Create an environment and install depenpencies

On Linux and Windows

  • To create CPU environment: conda env create -f release/env_cpu.yml
  • To create GPU environment: conda env create -f release/env_gpu.yml

On macOS

  • To create CPU environment: conda env create -f release/env_osx.yml

💥 Quick test

Step 1: Download models and pretrained checkpoints into release/models directory

Step 2: Download sample images and extract into release/sample_images directory

stefann/
├── ...
├── release/
│   ├── models/
│   │   ├── colornet.json
│   │   ├── colornet_weights.h5
│   │   ├── fannet.json
│   │   └── fannet_weights.h5
│   ├── sample_images/
│   │   ├── 01.jpg
│   │   ├── 02.jpg
│   │   └── ...
│   └── ...
└── ...

Step 3: Activate environment

To activate CPU environment: conda activate stefann-cpu
To activate GPU environment: conda activate stefann-gpu

Step 4: Change directory to release and run STEFANN

cd release
python stefann.py

2. Editing Results 😆


Each image pair consists of the original image (Left) and the edited image (Right).


Training Networks

1. Downloading Datasets

Download datasets and extract the archives into datasets directory under repository root.

stefann/
├── ...
├── datasets/
│   ├── fannet/
│   │   ├── pairs/
│   │   ├── train/
│   │   └── valid/
│   └── colornet/
│       ├── test/
│       ├── train/
│       └── valid/
└── ...

📌 Description of datasets/fannet

This dataset is used to train FANnet and it consists of 3 directories: fannet/pairs, fannet/train and fannet/valid. The directories fannet/train and fannet/valid consist of 1015 and 300 sub-directories respectively, each corresponding to one specific font. Each font directory contains 64x64 grayscale images of 62 English alphanumeric characters (10 numerals + 26 upper-case letters + 26 lower-case letters). The filename format is xx.jpg where xx is the ASCII value of the corresponding character (e.g. "48.jpg" implies an image of character "0"). The directory fannet/pairs contains 50 image pairs, each corresponding to a random font from fannet/valid. Each image pair is horizontally concatenated to a dimension of 128x64. The filename format is id_xx_yy.jpg where id is the image identifier, xx and yy are the ASCII values of source and target characters respectively (e.g. "00_65_66.jpg" implies a transformation from source character "A" to target character "B" for the image with identifier "00").

📌 Description of datasets/colornet

This dataset is used to train Colornet and it consists of 3 directories: colornet/test, colornet/train and colornet/valid. Each directory consists of 5 sub-directories: _color_filters, _mask_pairs, input_color, input_mask and output_color. The directory _color_filters contains synthetically generated color filters of dimension 64x64 including both solid and gradient colors. The directory _mask_pairs contains a set of 64x64 grayscale image pairs selected at random from 1315 available fonts in datasets/fannet. Each image pair is horizontally concatenated to a dimension of 128x64. For colornet/train and colornet/valid each color filter is applied on each mask pair. This results in 64x64 image triplets of color source image, binary target image and color target image in input_color, input_mask and output_color directories respectively. For colornet/test one color filter is applied only on one mask pair to generate similar image triplets. With a fixed set of 100 mask pairs, 80000 colornet/train and 20000 colornet/valid samples are generated from 800 and 200 color filters respectively. With another set of 50 mask pairs, 50 colornet/test samples are generated from 50 color filters.

2. Training FANnet and Colornet

Step 1: Activate environment

To activate CPU environment: conda activate stefann-cpu
To activate GPU environment: conda activate stefann-gpu

Step 2: Change directory to project root

cd stefann

Step 3: Configure and train FANnet

To configure training options edit configurations section (line 40-72) of fannet.py
To start training: python fannet.py

☁️ Check this notebook hosted at Kaggle for an interactive demonstration of FANnet.

Step 4: Configure and train Colornet

To configure training options edit configurations section (line 38-65) of colornet.py
To start training: python colornet.py

☁️ Check this notebook hosted at Kaggle for an interactive demonstration of Colornet.

External Links

Project  •   Paper  •   Supplementary Materials  •   Datasets  •   Models  •   Sample Images


Citation

@InProceedings{Roy_2020_CVPR,
  title     = {STEFANN: Scene Text Editor using Font Adaptive Neural Network},
  author    = {Roy, Prasun and Bhattacharya, Saumik and Ghosh, Subhankar and Pal, Umapada},
  booktitle = {The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  month     = {June},
  year      = {2020}
}

License

Copyright 2020 by the authors

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
Made with ❤️ and 🍕 on Earth.
This is a pytorch re-implementation of EAST: An Efficient and Accurate Scene Text Detector.

EAST: An Efficient and Accurate Scene Text Detector Description: This version will be updated soon, please pay attention to this work. The motivation

Dejia Song 544 Dec 20, 2022
Detect textlines in document images

Textline Detection Detect textlines in document images Introduction This tool performs border, region and textline detection from document image data

QURATOR-SPK 70 Jun 30, 2022
Give a solution to recognize MaoYan font.

猫眼字体识别 该 github repo 在于帮助xjtlu的同学们识别猫眼的扭曲字体。已经打包上传至 pypi ,可以使用 pip 直接安装。 猫眼字体的识别不出来的原理与解决思路在采茶上 使用方法: import MaoYanFontRecognize

Aruix 4 Jun 30, 2022
Détection de créneaux de vaccination disponibles pour l'outil ViteMaDose

Vite Ma Dose ! est un outil open source de CovidTracker permettant de détecter les rendez-vous disponibles dans votre département afin de vous faire v

CovidTracker 239 Dec 13, 2022
document image degradation

ocrodeg The ocrodeg package is a small Python library implementing document image degradation for data augmentation for handwriting recognition and OC

NVIDIA Research Projects 134 Nov 18, 2022
Camera Intrinsic Calibration and Hand-Eye Calibration in Pybullet

This repository is mainly for camera intrinsic calibration and hand-eye calibration. Synthetic experiments are conducted in PyBullet simulator. 1. Tes

CAI Junhao 7 Oct 03, 2022
QuanTaichi: A Compiler for Quantized Simulations (SIGGRAPH 2021)

QuanTaichi: A Compiler for Quantized Simulations (SIGGRAPH 2021) Yuanming Hu, Jiafeng Liu, Xuanda Yang, Mingkuan Xu, Ye Kuang, Weiwei Xu, Qiang Dai, W

Taichi Developers 119 Dec 02, 2022
An advanced 2D image manipulation with features such as edge detection and image segmentation built using OpenCV

OpenCV-ToothPaint3-Advanced-Digital-Image-Editor This application named ‘Tooth Paint’ version TP_2020.3 (64-bit) or version 3 was developed within a w

JunHong 1 Nov 05, 2021
Handwritten Number Recognition using CNN and Character Segmentation

Handwritten-Number-Recognition-With-Image-Segmentation Info About this repository This Repository is aimed at reading handwritten images of numbers an

Sparsha Saha 17 Aug 25, 2022
Single Shot Text Detector with Regional Attention

Single Shot Text Detector with Regional Attention Introduction SSTD is initially described in our ICCV 2017 spotlight paper. A third-party implementat

Pan He 215 Dec 07, 2022
PSENet - Shape Robust Text Detection with Progressive Scale Expansion Network.

News Python3 implementations of PSENet [1], PAN [2] and PAN++ [3] are released at https://github.com/whai362/pan_pp.pytorch. [1] W. Wang, E. Xie, X. L

1.1k Dec 24, 2022
PianoVisuals - Create background videos synced with piano music using opencv

Steps Record piano video Use Neural Network to do body segmentation (video matti

Solbiati Alessandro 4 Jan 24, 2022
~1000 book pages + OpenCV + python = page regions identified as paragraphs, lines, images, captions, etc.

cosc428-structor I had an open-ended Computer Vision assignment to complete, and an out-of-copyright book that I wanted to turn into an ebook. Convent

Chad Oliver 45 Dec 06, 2022
governance proposal to make fei redeemable for eth

Feil Proposal 🌲 Abstract Migrate all ETH from Fei protocol-controlled value into Yearn ETH Vault. Allow redemptions of outstanding FEI for yvETH. At

13 Mar 31, 2022
Code release for our paper, "SimNet: Enabling Robust Unknown Object Manipulation from Pure Synthetic Data via Stereo"

SimNet: Enabling Robust Unknown Object Manipulation from Pure Synthetic Data via Stereo Thomas Kollar, Michael Laskey, Kevin Stone, Brijen Thananjeyan

68 Dec 14, 2022
Image augmentation library in Python for machine learning.

Augmentor is an image augmentation library in Python for machine learning. It aims to be a standalone library that is platform and framework independe

Marcus D. Bloice 4.8k Jan 04, 2023
Dataset and Code for ICCV 2021 paper "Real-world Video Super-resolution: A Benchmark Dataset and A Decomposition based Learning Scheme"

Dataset and Code for RealVSR Real-world Video Super-resolution: A Benchmark Dataset and A Decomposition based Learning Scheme Xi Yang, Wangmeng Xiang,

Xi Yang 91 Nov 22, 2022
Histogram specification using openCV in python .

histogram specification using openCV in python . Have to input miu and sigma to draw gausssian distribution which will be used to map the input image . Example input can be miu = 128 sigma = 30

Tamzid hasan 6 Nov 17, 2021
Document Image Dewarping

Document image dewarping using text-lines and line Segments Abstract Conventional text-line based document dewarping methods have problems when handli

Taeho Kil 268 Dec 23, 2022
Demo for the paper "Overlap-aware low-latency online speaker diarization based on end-to-end local segmentation"

Streaming speaker diarization Overlap-aware low-latency online speaker diarization based on end-to-end local segmentation by Juan Manuel Coria, Hervé

Juanma Coria 185 Jan 01, 2023