Zum Hauptinhalt springen

Import images

Access the Dataset tab to manage training data.

Training Data View

Import Methods​

MethodDescription
Drag & DropImport image files directly into the workspace
Import Files / Import FoldersLoad unlabeled image data from the local filesystem
Import DatasetLoad a pre-labeled dataset in a supported annotation format
Camera ToolCapture images directly from connected cameras (multi-camera supported)
Spectrogram GeneratorConvert audio or CSV files into spectrogram images

Supported Annotation Formats​

Dataset Import

FormatDescription
YOLO.txt annotation files with normalized bounding box coordinates
COCOJSON annotation file with bounding boxes and segmentation masks
Pascal VOCXML annotation files per image
ClassificationDirectory-per-class folder structure
PNG Mask (Semantic)Per-image PNG masks where pixel values encode class IDs — for semantic segmentation datasets
Example Datasets

Try one of these demo datasets to get started quickly: