Generates synthetic datasets for multi-line intersection construction. The agent must extend three given line segments into full lines, observe their common intersection point, and identify which of several option labels coincides with it — testing geometric extrapolation and concurrence reasoning.
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 | Multi-26 |
| Task | Line Intersection Construction |
| Category | Constructive Geometry |
| Resolution | 1024×1024 px |
| FPS | 16 fps |
| Duration | varies |
| Output | PNG images + MP4 video |
# 1. Clone the repository
git clone https://github.com/VBVR-DataFactory/Multi-26_line_intersection_construction_data-generator.git
cd Multi-26_line_intersection_construction_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
# Reproducible generation with seed
python examples/generate.py --num-samples 50 --seed 42
# Custom output directory
python examples/generate.py --num-samples 100 --output data/my_dataset
# Without videos (faster, images only)
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) |
[Scenario] The image displays three separate line segments and several option markers (letters).
[Rules]
1. The three line segments must be extended into infinite straight lines.
2. The three lines will intersect at one single common point.
[Task] Generate a video animating the extension of the three line segments into full lines. Keep the extensions visible, identify their common intersection point, and highlight in red the option marker located at that specific point. Hold the final frame.
 |
 |
 |
| Initial Frame Three short segments + option labels |
Animation Each segment extended into a full line |
Final Frame Common intersection found; correct option highlighted |
Extend three given short line segments into infinite lines along their respective directions, identify the single point where all three lines meet (concurrence point), and select which option label is located at that point.
- Three segments: Short, randomly positioned and oriented line segments.
- Concurrence: Generated such that the three full lines meet at exactly one common point.
- Option labels: A, B, C, D candidate locations placed near (but not coincident with) the intersection.
- Solver: Closed-form line-line intersection in homogeneous coordinates.
- Extrapolation reasoning: The intersection lies outside the visible segments, requiring directional extrapolation, not interpolation.
- Three-way concurrence: A correct answer must coincide with all three lines, eliminating two-line-only false positives.
- Quantitative discrimination: Distractor option labels are placed close to the true intersection, demanding precise extension.
- Visible construction: Each segment's extension renders as a separate animation phase, giving stepwise correctness signal.
data/questions/Multi-26_line_intersection_construction_data-generator_task/Multi-26_line_intersection_construction_data-generator_00000000/
├── first_frame.png # Three short segments + option labels
├── final_frame.png # Lines extended + correct option marked
├── prompt.txt # Task instruction
├── ground_truth.mp4 # Animation of segment extensions
└── question_metadata.json # Standardized VBVR task metadata
File specifications:
- Images: 1024×1024 PNG format
- Video: MP4 format, 16 fps, H.264 + yuv420p
- Metadata: VBVR canonical schema with
task_id,vbvr_task_code,media,parameters
line-intersection concurrence geometric-construction extrapolation multiple-choice multi-step-reasoning
Part of the 36-Task Long-Horizon Multi-Step Reasoning Benchmark.