surawut/chessboard-dataset-yolo

---
language:
- en
license: cc-by-4.0
task_categories:
- keypoint-detection
tags:
- chess
- chessboard
- yolo
- pose-estimation
- corner-detection
- computer-vision
size_categories:
- 1K<n<10K
pretty_name: Chessboard Corner Pose Dataset
---

# Chessboard Corner Pose Dataset ♟️📐

## 1. Dataset Description
This dataset is designed for **Keypoint Detection (Pose Estimation)** tasks, specifically to detect the 4 semantic corners of a chessboard from various camera angles.

It is intended to be used as **Phase 1** of a Chess Move Tracking pipeline. The detected keypoints are used to calculate the Homography matrix for Perspective Warping (straightening the board).

- **Task:** Keypoint Detection (YOLO-Pose)
- **Format:** YOLOv8 / YOLO11 Pose Format
- **Input:** Raw video frames (rotated, tilted, occluded)
- **Output:** 4 Keypoints (`a1`, `h1`, `a8`, `h8`)

## 2. Dataset Structure

### Keypoints Definition
The model is trained to detect **1 Class (Chessboard)** with **4 Keypoints**:
1. **Index 0:** `a1` (Bottom-Left in White perspective)
2. **Index 1:** `h1` (Bottom-Right in White perspective)
3. **Index 2:** `a8` (Top-Left in White perspective)
4. **Index 3:** `h8` (Top-Right in White perspective)

### Data Splits
| Split | Images | Description |
| :--- | :--- | :--- |
| **Train** | 3390 | Extracted from Various Chess Play Videos with augmentations |
| **Valid** | 290 | Extracted from Various Chess Play Videos (Unseen angles) |