Lexski/sudoku-image-recognition
收藏资源简介:
该数据集包含1400张标记的数独谜题图像,用于训练和评估能够自动确定谜题中每个单元格状态的系统。图像分为训练集(1000张)、验证集(200张)和测试集(200张)。数据集的结构包括图像、单元格数组和关键点数组。图像格式多样,包括webp、jpg和png,部分为截图,部分为照片。单元格数组表示谜题单元格的状态,关键点数组表示谜题的关键点坐标。
This dataset consists of 1400 labelled images of Sudoku puzzles for multilabel classification. It is intended for training and evaluating a system that can automatically determine the state of each cell in the puzzle: whether it is solved or unsolved, and which digits it contains. The images are split into train (1000), val (200) and test (200). The dataset includes images in various formats and sizes, with annotations for puzzle cells and keypoints. The primary use of this dataset is to train models for automatically determining the state of a Sudoku puzzle from an image, which can enhance user experience in Sudoku applications.
数据集概述
许可证信息
- 许可证类型: MIT




