From b20e2dd53c4b9e17ec6a4510c63ed058444f056e Mon Sep 17 00:00:00 2001 From: zhoujun Date: Mon, 16 Jan 2023 19:44:44 +0800 Subject: [PATCH] add Data annotation link (#8862) * fixed opencv version * update infernence * add Data annotation link * update opencv version --- doc/doc_ch/table_recognition.md | 17 +++++++++++------ doc/doc_en/table_recognition_en.md | 5 +++++ requirements.txt | 4 ++-- 3 files changed, 18 insertions(+), 8 deletions(-) diff --git a/doc/doc_ch/table_recognition.md b/doc/doc_ch/table_recognition.md index f09dedd0..8b8dad66 100644 --- a/doc/doc_ch/table_recognition.md +++ b/doc/doc_ch/table_recognition.md @@ -6,6 +6,7 @@ - [1.1. 数据集格式](#11-数据集格式) - [1.2. 数据下载](#12-数据下载) - [1.3. 数据集生成](#13-数据集生成) + - [1.4 数据标注](#14-数据标注) - [2. 开始训练](#2-开始训练) - [2.1. 启动训练](#21-启动训练) - [2.2. 断点训练](#22-断点训练) @@ -39,15 +40,15 @@ img_label 每一行的json格式为: ```txt { - 'filename': PMC5755158_010_01.png, # 图像名 - 'split': ’train‘, # 图像属于训练集还是验证集 - 'imgid': 0, # 图像的index + 'filename': PMC5755158_010_01.png, # 图像名 + 'split': ’train‘, # 图像属于训练集还是验证集 + 'imgid': 0, # 图像的index 'html': { - 'structure': {'tokens': ['', '', '', ...]}, # 表格的HTML字符串 + 'structure': {'tokens': ['', '', '', ...]}, # 表格的HTML字符串 'cells': [ { - 'tokens': ['P', 'a', 'd', 'd', 'l', 'e', 'P', 'a', 'd', 'd', 'l', 'e'], # 表格中的单个文本 - 'bbox': [x0, y0, x1, y1] # 表格中的单个文本的坐标 + 'tokens': ['P', 'a', 'd', 'd', 'l', 'e', 'P', 'a', 'd', 'd', 'l', 'e'], # 表格中的单个文本 + 'bbox': [x0, y0, x1, y1] # 表格中的单个文本的坐标 } ] } @@ -78,6 +79,10 @@ TableGeneration是一个开源表格数据集生成工具，其通过浏览器 |简单表格|![](https://raw.githubusercontent.com/WenmuZhou/TableGeneration/main/imgs/simple.jpg)| |彩色表格|![](https://raw.githubusercontent.com/WenmuZhou/TableGeneration/main/imgs/color.jpg)| +## 1.4 数据标注 + +数据标注可参考[PPOCRLabel](../../PPOCRLabel/README_ch.md) + # 2. 开始训练 PaddleOCR提供了训练脚本、评估脚本和预测脚本，本节将以 [SLANet](../../configs/table/SLANet.yml) 模型训练PubTabNet英文数据集为例： diff --git a/doc/doc_en/table_recognition_en.md b/doc/doc_en/table_recognition_en.md index d79d9893..c0a1aa9d 100644 --- a/doc/doc_en/table_recognition_en.md +++ b/doc/doc_en/table_recognition_en.md @@ -6,6 +6,7 @@ This article provides a full-process guide for the PaddleOCR table recognition m - [1.1. DataSet Format](#11-dataset-format) - [1.2. Data Download](#12-data-download) - [1.3. Dataset Generation](#13-dataset-generation) + - [1.4 Data annotation](#14-data-annotation) - [2. Training](#2-training) - [2.1. Start Training](#21-start-training) - [2.2. Resume Training](#22-resume-training) @@ -80,6 +81,10 @@ Some samples are as follows: |Simple Table|![](https://raw.githubusercontent.com/WenmuZhou/TableGeneration/main/imgs/simple.jpg)| |Simple Color Table|![](https://raw.githubusercontent.com/WenmuZhou/TableGeneration/main/imgs/color.jpg)| +## 1.4 Data annotation + +Data annotation can refer to[PPOCRLabel](../../PPOCRLabel/README.md) + # 2. Training PaddleOCR provides training scripts, evaluation scripts, and prediction scripts. In this section, the [SLANet](../../configs/table/SLANet.yml) model will be used as an example: diff --git a/requirements.txt b/requirements.txt index d3484033..21a5c463 100644 --- a/requirements.txt +++ b/requirements.txt @@ -7,8 +7,8 @@ tqdm numpy visualdl rapidfuzz -opencv-python==4.6.0.66 -opencv-contrib-python==4.6.0.66 +opencv-python<=4.6.0.66 +opencv-contrib-python<=4.6.0.66 cython lxml premailer -- GitLab