Browse categories
Explore
Fiverr Pro
English
$
USD
I will do image annotation, bounding box, and polygon segmentation
About this Gig
Need accurate image annotation for your machine learning or computer vision project? I provide precise manual labeling so your models get clean, ready-to-use datasets.
Whether you're building object detection pipelines or need detailed segmentation masks, I take care of the tedious labeling work so you can focus on training your models.
What I Can Do For You:
- Bounding Boxes: Clean boxes for object detection (cars, people, products, etc.).
- Polygon Segmentation: Precise dot-to-dot outlines for irregular shapes or detailed objects.
- Keypoint & Polyline Tagging: Point marking for pose tracking or line tracing for lanes and paths.
Formats I Deliver:
- YOLO (.txt)
- COCO (.json)
- Pascal VOC (.xml)
- CSV or Mask PNGs
Tools I Work With: CVAT, Roboflow, LabelImg, and Label Studio.
Why Work With Me?
- 100% Manual Work: No lazy auto-labeling hacks. Every box and polygon is drawn by hand.
- Pixel Accuracy: Tight boundaries with no loose margins or missed objects.
- Flexible Revisions: If something isn't labeled to your exact specs, I'll fix it right away.
Note: Please send me a quick message before ordering with a sample image and your label list so we can confirm the format and scope!
Technique:
Manual
Tagging type:
Image
•
Video
FAQ
Q: What information do I need to provide before starting the project?
A: Please provide your raw images, a clear list of object classes to label, any specific guidelines (e.g., how to handle partially hidden objects), and your preferred export format.
Q: Which annotation formats can you deliver?
A: I can export your annotations in all major machine learning formats, including YOLO (.txt), COCO (.json), Pascal VOC (.xml), CSV, and color-coded segmentation mask PNGs.
Q: What is the difference between Bounding Boxes and Polygon Segmentation?
A: Bounding boxes are rectangular boxes drawn around objects, best suited for fast object detection models. Polygon segmentation involves tracing the exact pixel-level outline of an object, which is ideal for complex shapes, medical images, or detailed instance segmentation.
