Generates synthetic datasets for training and evaluating vision models on chart reading and data comprehension tasks. Each sample contains a data table where specific values must be identified and highlighted based on semantic queries.
Each sample pairs a task (first frame + prompt describing what needs to happen) with its ground truth solution (final frame showing the result + video demonstrating how to achieve it). This structure enables both model evaluation and training.
| Property | Value |
|---|---|
| Task ID | G-27 |
| Task | Read The Chart Data Semantic Comprehension |
| Category | Knowledge |
| Resolution | 1024×1024 px |
| FPS | 16 fps |
| Duration | ~3 seconds |
| Output | PNG images + MP4 video |
# 1. Clone the repository
git clone https://github.com/VBVR-DataFactory/G-27_read_the_chart_data_semantic_comprehension_data-generator.git
cd G-27_read_the_chart_data_semantic_comprehension_data-generator
# 2. Create and activate virtual environment
python3 -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
# 3. Install dependencies
pip install --upgrade pip
pip install -r requirements.txt
pip install -e .# Generate 50 samples
python examples/generate.py --num-samples 50
# Custom output directory
python examples/generate.py --num-samples 100 --output data/my_dataset
# Reproducible generation with seed
python examples/generate.py --num-samples 50 --seed 42
# Without videos (faster)
python examples/generate.py --num-samples 50 --no-videos| Argument | Description |
|---|---|
--num-samples |
Number of tasks to generate (required) |
--output |
Output directory (default: data/questions) |
--seed |
Random seed for reproducibility |
--no-videos |
Skip video generation (images only) |
The scene shows a sales table with Class as rows and Goods as columns. Find the minimum value within the row corresponding to the Class 'Class 4' and draw a red rectangular border around the corresponding cell to highlight it.
 |
 |
 |
| Initial Frame Sales table with Areas and Products |
Animation Red border appears around target cell |
Final Frame Maximum value in West row highlighted |
Read and comprehend data tables to find specific values based on semantic queries (e.g., maximum in a specific row), then highlight the target cell.
- Data structure: Tables with rows and columns (e.g., Area × Product sales data)
- Query types: Find maximum/minimum values within specific rows or columns
- Row/column labels: Semantic names (e.g., "West", "Product A")
- Highlight method: Red rectangular border around target cell
- Data values: Numerical data in table cells
- Background: White with clear table structure
- Goal: Identify correct cell based on query and highlight it
- Semantic data comprehension (understanding row/column meanings)
- Maximum/minimum value identification within constraints
- Table structure reading and navigation
- Visual highlighting with red border annotation
- Tests data literacy and reading comprehension
- Various table configurations (different dimensions and labels)
data/questions/read_the_chart_data_semantic_comprehension_task/read_the_chart_data_semantic_comprehension_00000000/
├── first_frame.png # Data table without highlighting
├── final_frame.png # Table with target cell highlighted
├── prompt.txt # Query instruction specifying search criteria
├── ground_truth.mp4 # Animation of highlight appearing
└── question_metadata.json # Task metadata
File specifications:
- Images: 1024×1024 PNG format
- Video: MP4 format, 16 fps
- Duration: ~3 seconds
logic-symbols data-comprehension table-reading semantic-reasoning data-analysis visual-highlighting