Chessboard-digital-images_with_fen
收藏资源简介:
棋盘检测数据集包含总共64,386张棋盘图片及其对应的YOLO格式标签文件。数据集分为训练集和验证集,分别包含57,928和6,458张图片及标签。标签文件中包含棋盘上棋子的边界框,采用YOLO格式。数据集包含12个类别,包括6种白色棋子和6种黑色棋子。数据集是通过从Lichess平台获取的棋局数据生成的,使用python-chess API将FEN字符串转换为渲染的棋盘图片,并通过自定义脚本提取对象注释并转换为YOLO格式的.txt文件。该数据集适用于训练物体检测模型、检测棋盘上的单个棋子以及将棋盘图片转换回数字游戏状态(FEN)。数据集根据知识共享署名4.0国际(CC BY 4.0)许可证分发。
The chessboard detection dataset contains a total of 64,386 chessboard images and their corresponding YOLO-format label files. The dataset is split into training and validation sets, with 57,928 and 6,458 images and their corresponding labels, respectively. The label files contain bounding boxes of chess pieces on the chessboard, adhering to the YOLO format. The dataset includes 12 categories, encompassing 6 types of white chess pieces and 6 types of black chess pieces. This dataset is generated using game data collected from the Lichess platform. Specifically, the python-chess API is utilized to convert FEN strings into rendered chessboard images, while custom scripts are deployed to extract object annotations and convert them into YOLO-formatted .txt files. This dataset is applicable for training object detection models, detecting individual chess pieces on chessboards, and converting chessboard images back into digital game states (FEN). The dataset is distributed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.
Chessboard Detection Dataset 概述
基本描述
- 数据集名称: Chessboard Detection Dataset
- 语言: 英文 (en)
- 许可证: Creative Commons Attribution 4.0 International (CC BY 4.0)
- 任务类别: 目标检测 (object-detection)
- 任务ID: object-detection-yolo
- 规模分类: 10K<n<100K
数据集内容
- 总图像数量: 64,386
- 训练集 (train): 57,928
- 验证集 (val): 6,458
- 总标签数量: 64,386 (每个图像对应一个
.txt文件)- 训练集 (train): 57,928
- 验证集 (val): 6,458
标注信息
- 标注格式: YOLO格式
- 类别数量: 12
- 白色棋子: 6类
- 黑色棋子: 6类
数据来源与处理
- 数据来源: Lichess平台提供的国际象棋游戏数据(PGN格式)
- 处理工具:
python-chessAPI(用于将FEN字符串转换为棋盘图像)- 自定义脚本(用于分割棋盘并提取对象标注)
应用场景
- 训练目标检测模型(如YOLOv5, YOLOv8等)
- 检测棋盘上的单个棋子
- 将棋盘图像转换回数字游戏状态(FEN)
许可证信息
- 许可证类型: CC BY 4.0




