JieZi (解字)

JieZi (解字) is a large-scale, expert-audited visual question answering (VQA) dataset dedicated to ancient Chinese character exegesis. It pairs high-quality character glyph images with fine-grained expert annotations across nine paleographic tasks—including headword recognition, etymology, structural analysis, glyph evolution, and component function—providing a rigorous benchmark for multimodal understanding of Chinese paleography and calligraphy.

📊 Dataset at a Glance

Property	Value
Task	Visual Question Answering (VQA)
Language	Chinese (中文)
Total VQA Triplets	495,866
Unique Character Images	130,162
Unique Characters	8,762
Question Templates	7,208
Image Format	PNG / JPG
License	CC BY-NC-ND 4.0

Task Distribution

The dataset covers nine paleographic tasks. Each task asks a different dimension of question about the same set of glyph images:

Task	Chinese Name	Count	Percentage	Description
analysis	综合分析	130,162	26.3%	Comprehensive exegesis from form, meaning, and origin
headword	字头识别	45,713	9.2%	Identify the modern standard Chinese character (字头)
script	字形判断	45,713	9.2%	Classify the script type (甲骨文, 金文, 战国文字, 篆书, 隶书, 楷书, 草书)
etymology	造字法	45,713	9.2%	Determine the method of character creation
structure	结构分析	45,713	9.2%	Analyze the structural composition (e.g., 独体, 左右, 上下)
meaning	本义解释	45,713	9.2%	Explain the original meaning (本义) of the character
evolution	字形演变	45,713	9.2%	Describe the historical glyph evolution
components	构件分析	45,713	9.2%	Analyze each component's function and evolution
function	构件功能	45,713	9.2%	Identify the semantic/phonetic function of a given component

Script Distribution

Script	Count	Percentage
隶书 (Clerical)	117,450	23.7%
篆书 (Seal)	111,340	22.5%
草书 (Cursive)	98,846	19.9%
楷书 (Regular)	69,349	14.0%
甲骨文 (Oracle Bone)	59,269	12.0%
金文 (Bronze)	35,474	7.2%
战国文字 (Ancient)	4,138	0.8%

🗂️ Data Format

Each sample is a JSON object stored as one JSONL line. The full dataset has been shuffled and is ready for train/validation splitting.

{
  "image": "images/㐁_楷书_1.png",
  "question": "请识别图中的字形，并阐述其文字学属性。",
  "answer": "该字在现代字典中字头作「㐁」，图示为楷书字形。就造字法与结构而言，该字属独体结构之象形字...",
  "character": "㐁",
  "script": "楷书",
  "task": "analysis"
}

Field Description

Field	Type	Description
image	string	Relative path to the glyph image, rooted at JieZi-Dataset.jsonl. Format: images/{character}_{script}_{index}.{ext}. The filename embeds the character, script, and a sequential index for that glyph variant.
question	string	Natural-language question about the character. Multiple paraphrased templates exist for each task.
answer	string	Expert-verified ground-truth answer. Long-form for analysis, evolution, and components; short-form for headword, script, and etymology.
character	string	The modern standard simplified Chinese character (字头) that the depicted glyph corresponds to in contemporary dictionaries.
script	string	The script type (字形 / 书体): 甲骨文, 金文, 战国文字, 篆书, 隶书, 楷书, or 草书.
task	string	The task type: analysis, headword, script, etymology, structure, meaning, evolution, components, or function.

🔍 VQA Examples by Task

Below are representative samples for all nine task types. Each row shows the glyph image, the Chinese question, and the expert answer. Open the Dataset Viewer tab above to browse all 495,866 samples interactively.

1. Comprehensive Analysis (analysis)

Image	Question	Answer
	请对这幅古文字图像进行系统的释读。	观察图像所示甲骨文字形，该字在现代字典中字头为「雹」。从文字学角度分析，此字为⿱结构之象形字。其构件构成如下：「雨」承担表意功能，演变类型为保留，构字解释为「象天穹降雨之形，标示降水现象。」；「冰粒」承担表意功能，演变类型为讹变，构字解释为「象冰雹颗粒之形，后讹变为声符“包”。」...

2. Headword Recognition (headword)

Image	Question	Answer
	请确认图中字形对应的现代字典字头。	嶓

3. Script Classification (script)

Image	Question	Answer
	图中展示的是哪一种字形？	甲骨文

4. Etymology (etymology)

Image	Question	Answer
	这个字的六书造字法是什么？	形声字

5. Structural Analysis (structure)

Image	Question	Answer
	该字的结构应如何归类？	⿰

6. Original Meaning (meaning)

Image	Question	Answer
	图中字的原初意义（本义）是什么？	巡视田禾

7. Glyph Evolution (evolution)

Image	Question	Answer
	请依时代先后，说明该字字形的演变轨迹。	甲骨文作以戈击门形，金文讹变为手持棍敲门，后综合为从户从聿。篆文定型为三体结构，隶变后楷书分化为肁、肈、肇。草书在此基础上进一步连笔简化，保留上下结构特征。

8. Component Analysis (components)

Image	Question	Answer
	请完整分析该字的构件构成，包括每一构件的表意/表音功能、演变类型及解释。	1. 构件凵：功能=表意；演变类型=保留；解释=象坑穴或容器之形，在此表示关押野兽的笼槛。2. 构件牛：功能=表意；演变类型=保留；解释=象牛头及角之形，表示被关押的对象。

9. Component Function (function)

Image	Question	Answer
	请说明构件"水"在该字中的功能角色。	表意

🚀 Usage

Quick Start with Hugging Face datasets

from datasets import load_dataset

# Load from the Hub
dataset = load_dataset("Ran0/JieZi", split="train")

# Inspect a sample
sample = dataset[0]
print(sample["question"])   # 请识别图中的字形，并阐述其文字学属性。
print(sample["answer"])     # 该字在现代字典中字头作「㐁」...
print(sample["character"])  # 㐁 (modern standard character)
print(sample["script"])     # 楷书 (script type)
print(sample["task"])       # analysis (task type)
sample["image"].show()      # Display the character image

Manual Loading

If you prefer to work with the raw JSONL directly:

import json
from pathlib import Path

root = Path("JieZi-Dataset")
with (root / "JieZi-Dataset.jsonl").open("r", encoding="utf-8") as f:
    for line in f:
        item = json.loads(line)
        image_path = root / item["image"]   # relative path
        question   = item["question"]
        answer     = item["answer"]
        character  = item["character"]      # modern standard simplified Chinese character
        script     = item["script"]         # script type (e.g., 篆书, 隶书)
        task       = item["task"]           # task type (e.g., analysis, headword)
        # ... your processing logic ...

⚠️ Path Note: When using the raw JSONL, keep the relative path relationship between JieZi-Dataset.jsonl and images/ unchanged. Images are named as {character}_{script}_{index}.{ext}, making it easy to locate and manage specific glyph variants.

🏗️ Project Structure

JieZi/
├── JieZi-Dataset/
│   ├── JieZi-Dataset.jsonl    # Main dataset file (495,866 VQA records, shuffled)
│   └── images/                # Glyph image directory (130,162 images)
│       ├── 㐁_楷书_1.png
│       ├── 㐆_隶书_1.jpg
│       └── ...                # Naming: {character}_{script}_{index}.{ext}
├── README.md                  # Dataset documentation (this file)
└── main_figure.png            # Project teaser figure

📚 Citation

If you use this dataset in your research, please cite:

@misc{jiezi2024,
  title={JieZi: A Large-Scale Expert-Audited Dataset and Benchmark for Ancient Chinese Character Exegesis},
  author={Ran et al.},
  year={2024},
  howpublished={\url{https://huggingface.co/datasets/Ran0/JieZi}}
}

📜 License

This dataset is released under CC BY-NC-ND 4.0 for non-commercial research purposes only. By downloading or using the data, you agree to comply with the terms of this license.