YOLOSet: Data Factory

Initializing...

Generate Synthetic Dataset

Core Settings

Dataset Name

Total Number of Samples

Annotation

Task Type

Object Detection

Generates bounding boxes (YOLO).

Instance Segmentation

Generates polygon masks (YOLO).

Image Classification

Organizes images into class folders.

Segmentation Method

Convex Hull

Fastest

Concave Hull

Accurate

Render Mask

Slowest

Concavity 2 Lower values create more detailed shapes but can be slower. Higher values are faster and simpler.

Mask Simplification 0.1 Controls mask accuracy. Lower values create more detailed masks (more vertices). Higher values create simpler masks (fewer vertices).

Dataset Splits

Include Validation Set

Include Test Set

Split Ratios (100%)

Train

Val

Test

70%

20%

10%

Sample Augmentation

Include Background Samples Generate some images with no objects (negative samples). This helps reduce false positives in detection models.

Ratio of Backgrounds 15%

Scene & Randomization

3D Model Source

Use Current Model Randomize Library

Background Source

Use Current Panorama Randomize Library

Model Transform Applies random position, rotation, and scale. If 'Use Current Model' is selected, randomization is relative to its current state.

Position Var. 90%

Rotation Var. 100%

Scale Var. 10%

Camera Viewpoint Generate images from viewpoints similar to the current camera angle. Only available when using the current panorama.

Horizontal Var. 180°

Vertical Var. 20°

Generating Dataset

Processing...

Upload New Asset

Drag & Drop File

or click to browse