The Sensor AI Annotation Tool allows you to label images for computer vision tasks efficiently. You can annotate your own datasets or contribute to crowdsourced tasks to earn $SENSE tokens.
Step-by-Step Guide
Step 1: Upload Your Dataset
Log in to your Sensor AI account. Navigate to the "Annotation Tool" section. Click "Upload Dataset" and select your images (supports JPG, PNG, etc.). Step 2: Choose Annotation Type
Sensor AI supports multiple annotation formats:
Bounding Boxes (Rectangles) – For object detection. Polygons – For precise segmentation. Keypoints (Points) – For pose estimation. Lines – For lane detection (e.g., autonomous vehicles). Select the appropriate tool from the left-side toolbar.
Step 3: Start Labeling
Click and drag to draw shapes around objects. Assign class labels (e.g., "car," "pedestrian"). Use AI-assisted labeling (if available) to speed up the process: Click "AI Suggest" to auto-detect objects (powered by YOLOv5 or PoseNet). Manually refine any incorrect suggestions. Step 4: Export Annotations
Once labeling is complete:
Choose your preferred format: COCO JSON (for segmentation tasks). YOLO TXT (for Darknet/YOLO models). VOC XML (for PASCAL VOC compatibility). Download the file for training or upload it to the Dataset Marketplace. Pro Tips
✅ Use keyboard shortcuts (e.g., Ctrl+Z to undo).
✅ Enable "Auto-Save" to prevent data loss.
✅ Join crowdsourced annotation tasks to earn $SENSE tokens.
