level2-data-annotation_cv-level2-cv-15 created by GitHub Classroom

Overview

[AI Tech 3기 Level2 P Stage] 글자 검출 대회

image

팀원 소개

김규리_T3016 박정현_T3094 석진혁_T3109 손정균_T3111 이현진_T3174 임종현_T3182

Overview

OCR (Optimal Character Recognition) 기술은 사람이 직접 쓰거나 이미지 속에 있는 문자를 얻은 다음 이를 컴퓨터가 인식할 수 있도록 하는 기술로, 컴퓨터 비전 분야에서 현재 널리 쓰이는 대표적인 기술 중 하나입니다.

OCR task는 글자 검출 (text detection), 글자 인식 (text recognition), 정렬기 (Serializer) 등의 모듈로 이루어져 있는데 본 대회는 글자 검출 (text detection)만을 해결하게 됩니다.

데이터를 구성하고 활용하는 방법에 집중하는 것을 장려하는 취지에서, 제공되는 베이스 코드 중 모델과 관련한 부분을 변경하는 것이 금지되어 있습니다. 데이터 수집과 preprocessing, data augmentation 그리고 optimizer, learning scheduler 등 최적화 방식을 변경할 수 있습니다.

  • Input : 글자가 포함된 전체 이미지
  • Output : bbox 좌표가 포함된 UFO Format

평가방법

  • DetEval

    이미지 레벨에서 정답 박스가 여러개 존재하고, 예측한 박스가 여러개가 있을 경우, 박스끼리의 다중 매칭을 허용하여 점수를 주는 평가방법 중 하나 입니다

    1. 모든 정답/예측박스들에 대해서 Area Recall, Area Precision을 미리 계산해냅니다.

    2. 모든 정답 박스와 예측 박스를 순회하면서, 매칭이 되었는지 판단하여 박스 레벨로 정답 여부를 측정합니다.

    3. 모든 이미지에 대하여 Recall, Precision을 구한 이후, 최종 F1-Score은 모든 이미지 레벨에서 측정 값의 평균으로 측정됩니다.

      image

Final Score  🏅

  • Public : f1 0.6897 → Private f1 : 0.6751
  • Public : 11위/19팀 → Private : 9위/19팀

image

Archive contents

template
├──code
│  ├──augmentation.py
│  ├──convert_mlt.py
│  ├──dataset.py
│  ├──deteval.py
│  ├──east_dataset.py
│  ├──inference.py
│  ├──loss.py
│  ├──model.py
│  └──train.py
└──input
   └──ICDAR2017_Korean
		  └──data
			  	├──images
		      └──ufo
			        ├──train.json
							└──val.json

Dataset

  • ICDAR MLT17 Korean : 536 images ⊆ ICDAR MLT17 : 7,200 images

  • ICDAR MLT19 : 10,000 images

  • ICAR ArT : 5,603 images

Experiment

Results

dataset 데이터 수 LB score (public→private) Recall Precision
01 ICDAR17_Korean 536 0.4469 → 0.4732 0.3580 → 0.3803 0.5944 → 0.6264
02 Camper (폴리곤 수정 전) 1288 0.4543 → 0.5282 0.3627 → 0.4349 0.6077 → 0.6727
03 Camper (폴리곤 수정 후) 1288 0.4644 → 0.5298 0.3491 → 0.4294 0.6936 → 0.6913
04 ICDAR17_Korean + Camper 1824 0.4447 → 0.5155 0.3471 → 0.4129 0.6183 → 0.6858
05 ICDAR17(859) 859 0.5435 → 0.5704 0.4510 → 0.4713 0.6837 → 0.7222
06 ICDAR17_MLT 7200 0.6749 → 0.6751 0.5877 → 0.5887 0.7927 → 0.7912
07 ICDAR19+ArT 약 15000 0.6344 → 0.6404 0.5489 → 0.5607 0.7514 → 0.7465

Requirements

pip install -r requirements.txt

UFO Format으로 변환

python convert_mlt.py

SRC_DATASET_DIR = {변환 전 data 경로}

DST_DATASET_DIR = {변환 된 data 경로}

UFO Format ****

File Name
    ├── img_h
    ├── img_w
    └── words
        ├── points
        ├── transcription
        ├── language
        ├── illegibillity
        ├── orientation
        └── word_tags

Train.py

python train.py --data_dir {train data path} --val_data_dir {val data path} --name {wandb run name} --exp_name {model name
🐱‍🏍 A curated list of awesome things related to Hugo themes.

awesome-hugo-themes Automated deployment @ 2021-10-12 06:24:07 Asia/Shanghai &sorted=updated Theme Author License GitHub Stars Updated Blonde wamo MIT

13 Dec 12, 2022
Python code for working with NFL play by play data.

nfl_data_py nfl_data_py is a Python library for interacting with NFL data sourced from nflfastR, nfldata, dynastyprocess, and Draft Scout. Includes im

82 Jan 05, 2023
Example Python code for running the mango-explorer marketmaker

🥭 Mango Explorer 📖 Introduction This guide will show you how to load and run a customisable marketmaker that runs on Mango Markets using the mango-e

Blockworks Foundation 2 Apr 11, 2022
Tutorial for STARKs with supporting code in python

stark-anatomy STARK tutorial with supporting code in python Outline: introduction overview of STARKs basic tools -- algebra and polynomials FRI low de

121 Jan 03, 2023
This is a template (starter kit) for writing Maturity Work with Sphinx / LaTeX at Collège du Sud

sphinx-tm-template Ce dépôt est un template de base utilisable pour écrire ton travail de maturité dans le séminaire d'informatique du Collège du Sud.

6 Dec 22, 2022
step by step guide for beginners for getting started with open source

Step-by-Step Guide for beginners for getting started with Open-Source Here The Contribution Begins 💻 If you are a beginner then this repository is fo

Arpit Jain 66 Jan 03, 2023
Python Programming (Practical) (1-25) Download 👇🏼

BCA-603 : Python Programming (Practical) (1-25) Download zip 🙂 🌟 How to run programs : Clone or download this repo to your computer. Unzip (If you d

Milan Jadav 2 Jun 02, 2022
This is the repository that includes the code material for the ESweek 2021 for the Education Class Lecture A3 "Learn to Drive (and Race!) Autonomous Vehicles"

ESweek2021_educationclassA3 This is the repository that includes the code material for the ESweek 2021 for the Education Class Lecture A3 "Learn to Dr

F1TENTH Autonomous Racing Community 29 Dec 06, 2022
Xanadu Quantum Codebook is an experimental, exercise-based introduction to quantum computing using PennyLane.

Xanadu Quantum Codebook The Xanadu Quantum Codebook is an experimental, exercise-based introduction to quantum computing using PennyLane. This reposit

Xanadu 43 Dec 09, 2022
Quilt is a self-organizing data hub for S3

Quilt is a self-organizing data hub Python Quick start, tutorials If you have Python and an S3 bucket, you're ready to create versioned datasets with

Quilt Data 1.2k Dec 30, 2022
freeCodeCamp Scientific Computing with Python Project for Certification.

Polygon_Area_Calculator freeCodeCamp Python Project freeCodeCamp Scientific Computing with Python Project for Certification. In this project you will

Rajdeep Mondal 1 Dec 23, 2021
Żmija is a simple universal code generation tool.

Żmija Żmija is a simple universal code generation tool. It is intended to be used as a means to generate code that is both efficient and easily mainta

Adrian Samoticha 2 Nov 23, 2021
Python Tool to Easily Generate Multiple Documents

Python Tool to Easily Generate Multiple Documents Running the script doesn't require internet Max Generation is set to 10k to avoid lagging/crashing R

2 Apr 27, 2022
The source code that powers readthedocs.org

Welcome to Read the Docs Purpose Read the Docs hosts documentation for the open source community. It supports Sphinx docs written with reStructuredTex

Read the Docs 7.4k Dec 25, 2022
Minimal reproducible example for `mkdocstrings` Python handler issue

Minimal reproducible example for `mkdocstrings` Python handler issue

Hayden Richards 0 Feb 17, 2022
Count the number of lines of code in a directory, minus the irrelevant stuff

countloc Simple library to count the lines of code in a directory (excluding stuff like node_modules) Simply just run: countloc node_modules args to

Anish 4 Feb 14, 2022
Read write method - Read files in various types of formats

一个关于所有格式文件读取的方法 1。 问题描述: 各种各样的文件格式,读写操作非常的麻烦,能够有一种方法,可以整合所有格式的文件,方便用户进行读取和写入。 2

2 Jan 26, 2022
Project documentation with Markdown.

MkDocs Project documentation with Markdown. View the MkDocs documentation. Project release notes. Visit the MkDocs wiki for community resources, inclu

MkDocs 15.6k Jan 02, 2023
Beautiful static documentation generator for OpenAPI/Swagger 2.0

Spectacle The gentleman at REST Spectacle generates beautiful static HTML5 documentation from OpenAPI/Swagger 2.0 API specifications. The goal of Spec

Sourcey 1.3k Dec 13, 2022
Dev Centric Tools for Mkdocs Based Documentation

docutools MkDocs Documentation Tools For Developers This repo is providing a set of plugins for mkdocs material compatible documentation. It is meant

Axiros GmbH 14 Sep 10, 2022