Image Segmentation BASICS
Learn the conventions of pixel-level work. Semantic, instance, and panoptic segmentation for medical, AV, and satellite AI.
Self-Paced · Always Open
Free
Optional certificate: S$9
Overview
Start your Image Segmentation journey.
If image labeling is “draw a box around the dog”, image segmentation is “trace the exact outline of the dog, every pixel”. It’s slower, more precise, and the foundation of the work that pays better: medical imaging, autonomous driving perception, satellite analysis, microscopy. The skill at the heart of it is patience with detail and an eye for where exactly an object stops being itself.
This Basics course covers the conventions: semantic segmentation, instance segmentation, panoptic segmentation, and polygon and mask annotation. Self-paced theory and knowledge checks; no fixed schedule.
Part of the Data Annotation Basics learning track. Optional verifiable certificate available for a small fee.
What You Get
What's included in your training
Structured Learning Modules
Learn at your own pace through structured modules on the DeeLab Academy learning platform.
Self-Paced Schedule
No fixed schedule. Start when you want, finish at your own pace.
Knowledge Checks
Worked examples and per-unit practice quizzes with instant feedback, plus a final knowledge assessment on the e-learning platform (80% to pass).
Optional Certificate
Optional verifiable certificate available for a small fee after you pass the final assessment.
Outline
What you'll learn
-
Semantic Segmentation
How semantic segmentation labels every pixel by class (sky, road, person, car).
-
Instance Segmentation
What instance segmentation distinguishes: each separate object, not just its class.
-
Panoptic Segmentation
How panoptic segmentation combines semantic and instance labels into a unified map.
-
Polygon vs Mask Tradeoffs
When to choose polygon annotation versus mask annotation, and why precision matters.
-
Spotting Correct vs Incorrect Segmentation
How to recognise the difference between a clean contour and a sloppy one.
Part of the Data Annotation Basics programme.
Self-Paced Course
Free
Optional certificate: S$9
- Full access to the learning environment
- Complete at your own pace
- Professional certification optional
How our free courses work
🌱 One free course at a time — finish this one before you start another.
✍️ When you complete it, a quick rating and honest review are required before your next free course unlocks — the only cost, and how we keep them free.
Frequently Asked Questions
Everything you need to know about this course.
It's designed for beginners. No AI knowledge or coding required. Some intuition about basic image annotation helps but isn't a prerequisite; the course covers what bounding boxes are and then moves on to pixel-level work from there.
Image labelling typically means drawing a rectangle around an object and assigning a class. Image segmentation means drawing the exact contour of the object: every pixel either belongs to the object or doesn't. It's slower, more precise, and the foundation of any application that needs to know where exactly something starts and stops. Medical imaging (tumours), autonomous driving (drivable surface), satellite analysis (land use), microscopy (cells).
Self-paced theory lessons and an online knowledge assessment on the e-learning platform. The Basics tier is focused on understanding the conventions of segmentation: semantic, instance, and panoptic segmentation, plus polygon and mask annotation. There is no hands-on work with annotation tools in this course; that comes in the paid Image Segmentation Essentials course.
The course is free, but the certificate is not — claiming it is optional and costs a small fee. It's verifiable proof that you've completed the self-paced course and passed the online knowledge assessment. Useful on a CV or LinkedIn profile as foundation evidence. It does not mean you've been trained to production-segmentation standards (especially medical, where domain knowledge is required); the paid Image Segmentation Essentials course is the hands-on path. Essentials-level courses make up the Certified Data Annotator (CDA) programme.
Two natural directions from here. To broaden your foundation, the Data Annotation Basics (DAB) programme has free courses on image, video, audio, text, and 3D point cloud, all at the same level as this one. To go deeper into pixel-level segmentation specifically, the paid Image Segmentation Essentials course is the next level up: online sessions with an industry trainer, access to the labeling tools used in production, ongoing support, and a graded final assessment. The Essentials-level courses make up the Certified Data Annotator (CDA) programme.
You have three months to complete the course, counted from the end date of your first course run. The deadline is fixed and doesn't reset if you move to a later run.
Yes — you're automatically enrolled in the run you choose, and you can move to another scheduled run of the same course up to twice from your dashboard. Your payment carries over and your completion deadline stays pinned to your first run's end date.
From the blog
Related articles
More on these topics
Related Programmes
