Segment 3D Data Without Labels
Use SegmentAnything 3D, region growing, Euclidean clustering, and RANSAC to split point clouds into meaningful parts with zero annotated data. 6 hands-on lessons. Real scans.
See unsupervised segmentation in action
From a raw point cloud to clustered, labelled output without a single annotated sample.
Automation target. Every method in the course runs without labelled training data.
Students trained worldwide across 80 countries.
Production experience distilled into structured, repeatable workflows.
You don’t have labelled data. You still need to segment.
Every segmentation tutorial assumes you already have a labelled dataset. Your reality looks different. You have a new LiDAR scan, a new site, a new client domain, and nobody labelled a single point. Training a PointNet++ model would cost months of annotation and six figures.
The real move is to skip training altogether. Foundation models like SAM plus classical clustering give you 80% of a supervised model’s quality with zero annotation. The engineers who master these unsupervised methods ship in days, not months.
Labelled 3D data is scarce and expensive. One annotated aerial scan can cost 50K+ EUR. Foundation models and clustering-based segmentation skip that cost entirely. You get to production 10x faster. This is the shortcut that changes everything.
What you’ll build
Six segmentation techniques. Zero annotated data required.
SegmentAnything 3D
Adapt Meta’s SAM foundation model to 3D point clouds. Project to 2D, segment, lift back to 3D. Works on scans you have never seen before.
Region growing
Classical region growing. Group points sharing similar normals and curvature. Still the best baseline for planar environments.
Euclidean clustering
Graph-theory based clustering using KD-trees. Split disconnected objects instantly. The workhorse for indoor and urban scenes.
RANSAC shape detection
RANSAC plane, cylinder, and sphere extraction. Find primitives in cluttered scenes with robust outlier rejection.
3D data labelling
Turn unsupervised output into a labelled dataset. Review, correct, export. Feed the result into supervised training when you need a custom model.
End-to-end pipeline
Combine foundation models with classical algorithms. Build a hybrid pipeline that handles whatever the client throws at you.
I built this course after watching clients burn through six-figure annotation budgets that produced models with worse performance than a well-tuned unsupervised baseline. Every technique in here comes from real projects where we skipped the labelling step and shipped faster.
How this course works
Focused, hands-on, and weekend-sized.
100% asynchronous
Access everything 24/7 on the LMS. Self-paced. Finish in a weekend.
Code-along projects
Complete Python source code for every method. Clone the repo, run it, adapt it to your data.
Real datasets
Indoor scans, urban LiDAR, industrial captures. Real noise and real class imbalance, not synthetic toys.
Production patterns
Memory-efficient code, batch processing, error handling. Ready to drop into a real pipeline.
Lifetime access
One payment, permanent access. Every update included.
Upgrade path
This module is part of the Segmentor OS. Your purchase applies as credit if you upgrade later.
This is a focused, single-topic course on unsupervised segmentation. If you want the complete segmentation stack (PointNet++, RandLA-Net, algorithmic forge, HITL labelling, deployed app), consider the Segmentor OS instead. Your purchase here applies as credit if you upgrade.
The Curriculum
6 lessons. From foundation models to a shipped unsupervised pipeline.
Prerequisites
This course requires basic Python skills and comfort with point cloud data.
- Python (mid-level): comfortable with classes, NumPy, virtual environments
- Hardware: 16 GB RAM recommended. GPU helpful for SAM but not required for classical methods
- Basic point cloud experience: understanding of XYZ coordinates, file formats, and Open3D helpful
No prior deep learning or foundation model experience required.
Adapt SAM to 3D point clouds. Project to 2D, segment with the foundation model, lift results back to 3D. Zero training required.
Two workhorse algorithms. Region growing for smooth surfaces. Euclidean clustering for disconnected objects.
RANSAC for plane, cylinder, and sphere extraction. Robust outlier rejection in cluttered scenes.
Combine foundation models with classical algorithms into a production pipeline. 3D data labelling output. Export to supervised training formats if you need one later.
Your instructor
Dr. Florent Poux
I’ve spent 12+ years in 3D geospatial: from field surveys with total stations to building AI systems for Fortune 500 companies. I published the O’Reilly book on 3D Data Science with Python. I’ve advised startups valued at over 15M EUR. I’ve held a professorship, taught at university, and led R&D for some of the largest organizations in the space.
I don’t teach syntax. I teach judgment. Every module is built around real decisions I’ve faced in production. Which neural renderer fits an industrial inspection job. How to architect a semantic pipeline that doesn’t choke on 500M points. When to use algorithmic methods and when to switch to deep learning.
What students say
Engineers and researchers working with 3D segmentation.
“Our team processed 200M points for a highway survey. Before this course, we were stuck at 10M with crashes. The memory management module saved us weeks of work.”
“I’m a surveyor with 20 years of field experience. This gave me the Python and AI skills to modernize our entire workflow. Best investment I’ve made in my career.”
“RANSAC plus DBSCAN on a 120M-point mining dataset — segmented, volumetric change detected, and report ready in one afternoon. That used to take a week.”
“This course is the foundation I wish I had when I started with LiDAR. The feature extraction module alone reshaped how I approach every project.”
“I trained my first production segmenter on industrial scans within a week. The labelling and class-imbalance modules saved me from rebuilding the dataset twice.”
“The region-growing plus learned-feature pipeline cut our manual classification work in half on a 600M-point urban dataset. We finally retired the spreadsheet.”
“What I needed was someone showing me how to evaluate a segmenter, not just train one. The metrics module is the part most courses skip and the part that matters.”
“I came in fluent in Python but lost on 3D. After Module 2 I was reading PointNet++ papers like a normal human. The intuition Florent builds is the missing link.”
“We replaced a brittle classical pipeline with a sparse-conv network thanks to this course. Production accuracy jumped from 78 to 94 percent on the same data.”
“The training loop debugging chapter was worth the price by itself. I now know why my models stop learning instead of guessing at hyperparameters.”
Get lifetime access
One payment. Every lesson, every update, every line of code.
3D Unsupervised Segmentation
Complete unsupervised module + source code + real datasets + lifetime updates
- 6 hands-on lessons (8+ hours)i
- Complete Python source code + datasets
- SegmentAnything 3D integration
- Region growing + clustering + RANSAC
- Lifetime access + all future updatesi
- 90-day results guaranteei
Zero-risk guarantee: If you don’t see real results within 90 days, full refund. No questions.
The complete ecosystem
3D AI Architect Program
The complete spatial AI curriculum, delivered in 3 tiers. Pick the depth that matches where you are — Foundations to get moving, Professional for the full OS stack, Ultimate for live access and priority support.
- 3D AI Acceleratori: 17 episodes in 6 acts
- 3D Course Libraryi: 24+ standalone courses
- All 4 OS courses (Professional & Ultimate tiers)
- Neurones 3D software access
- Monthly drop-in sessions with Dr. Poux (Ultimate)
- Spatial AI job and market intel
- Priority support + services access (Ultimate)
- 300+ hours of content
What you’re getting access to
Everything I’ve built over 12+ years, from land surveying in the field to advising 15M EUR startups, compressed into one curriculum you can start today. Delivered by the first QUALIOPI-certified 3D geospatial academy.
Every pipeline was battle-tested on Fortune 500 projects processing billions of points. You’re getting the real playbook, not theory.
Methods validated by peer-reviewed publications, the ISPRS scientific community, and 1,500+ academic citations. Not guesswork.
Built by someone who surveyed in the field, defended a PhD, advised funded startups, and shipped products to Fortune 500 clients.
I share more free content than most people put behind a paywall. That’s intentional. I want you to know exactly what you’re getting before you invest. This course is the concentrated, structured version of everything I know. No fluff. No filler. Just the production path.
Find the right path for you
From single courses to the complete ecosystem.
| Feature | Standalone Course | 3D Unsupervised Segmentation | Course Library | 3D AI Architecti | Enterprise |
|---|---|---|---|---|---|
| Courses included | 1 topic | 6 lessons | Full catalogi | 3 OS courses + Library (tiered) | Custom |
| Hours of content | 2-8h | 8+ hours | 150+ hours | 300+ hours (tiered) | Custom |
| Production source code | ✓ | ✓ | ✓ | ✓ | ✓ |
| Lifetime access | ✓ | ✓ | – | ✓ | ✓ |
| 3D AI Accelerator Tracki | – | – | – | ✓ | ✓ |
| Neurones 3D softwarei | – | – | – | ✓ | ✓ |
| Spatial AI job & market inteli | – | – | – | ✓ | ✓ |
| Monthly drop-in sessionsi | – | – | – | ✓ | ✓ |
| Priority support + services accessi | – | – | – | ✓ tiered | ✓ |
| Custom onboardingi | – | – | – | – | ✓ |
| Team licensing | – | – | – | – | ✓ |
| Price | €97 – €497 | €197 | €1,297 | Starts at €1,999 | On request |
Straight answers
Do I need prior deep learning experience?
No. You need basic Python and point cloud familiarity. I explain SAM and the foundation model concepts from scratch.
What hardware do I need?
Minimum 16 GB RAM. A CUDA GPU speeds up SAM inference but is not mandatory. Classical methods run fine on CPU.
How is this different from SAM tutorials online?
Online tutorials apply SAM to 2D images. This course adapts it to 3D point clouds via projection, then combines it with classical clustering for a hybrid pipeline that actually works on real scans.
Can I upgrade to the Segmentor OS later?
Yes. Your purchase applies as credit toward the Segmentor OS. Email me when you are ready to upgrade.
How long do I have access?
Lifetime. One payment, permanent access. Every future update included.
What’s the refund policy?
90 days. Apply the material. If you are not satisfied, email me for a full refund.
Will this replace supervised segmentation?
Not always. Unsupervised methods get you 80% of the way with zero labelling cost. For custom class taxonomies or maximum accuracy, you still need supervised training. This course teaches you when to pick which.
Not sure if this course fits?
If you have specific questions about how the curriculum applies to your role, your team’s needs, or your technical background, I’m happy to help you figure it out before you commit.
Book a 15-min call