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
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
Compare two CSV files for differences. Colorize the differences and align the columns.

pretty-csv-diff Compare two CSV files for differences. Colorize the differences and align the columns. Command-Line Example Command-Line Usage usage:

Devon 6 Dec 29, 2022
This tutorial will guide you through the process of self-hosting Polygon

Hosting guide This tutorial will guide you through the process of self-hosting Polygon Before starting Make sure you have the following tools installe

Polygon 2 Jan 31, 2022
A hack to run custom shell commands when building documentation on Read the Docs.

readthedocs-custom-steps A hack to run custom steps when building documentation on Read the Docs. Important: This module should not be installed outsi

Niklas Rosenstein 5 Feb 22, 2022
A swagger tool for tornado, using python to write api doc!

SwaggerDoc About A swagger tool for tornado, using python to write api doc! Installation pip install swagger-doc Quick Start code import tornado.ioloo

aaashuai 1 Jan 10, 2022
Ultimaker Cura 2 Mooraker Upload Plugin

Klipper & Cura - Cura2MoonrakerPlugin Allows you to upload Gcode directly from Cura to your Klipper-based 3D printer (Fluidd, Mainsailos etc.) using t

214 Jan 03, 2023
Documentation of the QR code found on new Austrian ID cards.

Austrian ID Card QR Code This document aims to be a complete documentation of the format used in the QR area on the back of new Austrian ID cards (Per

Gabriel Huber 9 Dec 12, 2022
Some custom tweaks to the results produced by pytkdocs.

pytkdocs_tweaks Some custom tweaks for pytkdocs. For use as part of the documentation-generation-for-Python stack that comprises mkdocs, mkdocs-materi

Patrick Kidger 4 Nov 24, 2022
The purpose of this project is to share knowledge on how awesome Streamlit is and can be

Awesome Streamlit The fastest way to build Awesome Tools and Apps! Powered by Python! The purpose of this project is to share knowledge on how Awesome

Marc Skov Madsen 1.5k Jan 07, 2023
MkDocs Plugin allowing your visitors to *File > Print > Save as PDF* the entire site.

mkdocs-print-site-plugin MkDocs plugin that adds a page to your site combining all pages, allowing your site visitors to File Print Save as PDF th

Tim Vink 67 Jan 04, 2023
Credit EDA Case Study Using Python

This case study aims to identify patterns which indicate if a client has difficulty paying their installments which may be used for taking actions such as denying the loan, reducing the amount of loa

Purvi Padliya 1 Jan 14, 2022
A module filled with many useful functions and modules in various subjects.

Usefulpy Check out the Usefulpy site Usefulpy site is not always up to date Download and Import download and install with with pip download usefulpyth

Austin Garcia 1 Dec 28, 2021
A repository of links with advice related to grad school applications, research, phd etc

A repository of links with advice related to grad school applications, research, phd etc

Shaily Bhatt 946 Dec 30, 2022
Fully reproducible, Dockerized, step-by-step, tutorial on how to mock a "real-time" Kafka data stream from a timestamped csv file. Detailed blog post published on Towards Data Science.

time-series-kafka-demo Mock stream producer for time series data using Kafka. I walk through this tutorial and others here on GitHub and on my Medium

Maria Patterson 26 Nov 15, 2022
Types that make coding in Python quick and safe.

Type[T] Types that make coding in Python quick and safe. Type[T] works best with Python 3.6 or later. Prior to 3.6, object types must use comment type

Contains 17 Aug 01, 2022
A Json Schema Generator

JSON Schema Generator Author : Eru Michael About A Json Schema Generator. This is a generic program that: Reads a JSON file similar to what's present

1 Nov 10, 2021
Spin-off Notice: the modules and functions used by our research notebooks have been refactored into another repository

Fecon235 - Notebooks for financial economics. Keywords: Jupyter notebook pandas Federal Reserve FRED Ferbus GDP CPI PCE inflation unemployment wage income debt Case-Shiller housing asset portfolio eq

Adriano 825 Dec 27, 2022
Mayan EDMS is a document management system.

Mayan EDMS is a document management system. Its main purpose is to store, introspect, and categorize files, with a strong emphasis on preserving the contextual and business information of documents.

3 Oct 02, 2021
Proyecto - Desgaste y rendimiento de empleados de IBM HR Analytics

Acceder al código desde Google Colab para poder ver de manera adecuada todas las visualizaciones y poder interactuar con ellas. Links de acceso: Noteb

1 Jan 31, 2022
swagger-codegen contains a template-driven engine to generate documentation, API clients and server stubs in different languages by parsing your OpenAPI / Swagger definition.

Master (2.4.25-SNAPSHOT): 3.0.31-SNAPSHOT: Maven Central ⭐ ⭐ ⭐ If you would like to contribute, please refer to guidelines and a list of open tasks. ⭐

Swagger 15.2k Dec 31, 2022