• €159 or 2 monthly payments of €79

3D RECONSTRUCTION EXPLORATION

A Deep Exploration of 3D Reconstruction. Advanced Engineers only.

* only accessible to Stereo Vision Engineers, purchase the Stereo Vision course first.

⚠️ Edgeneers: You've reached an Unexplored side of 3D Computer Vision...

Dear Computer Vision Engineer, if you're looking to get lost in the abyssal world of 3D Reconstruction, then you landed on the right page.

Here's a quick story:

Several months ago, I was browsing Twitter when I got completely stunned by a tweet from Tesla's team sharing the solidly grounded results of their 3D Reconstruction algorithm:

Why was I so shocked? This algorithm creates entire 3D scenes just with cameras. And it hasn't been long until I noticed 3D Reconstruction was taking over entire companies.

Like Apple and their recently released "Object Capture API":

It didn't take me much longer to realize that most phone companies were trying to implement 3D Reconstruction using the 2 cameras of their devices, that the healthcare industry was implementing 3D reconstruction before surgeries, and even that the construction field was implementing these technologies.

Today, 3D Reconstruction is silently sneaking into our daily lives, whether it's healthcare, robotics, self-driving cars, drones, or any other field that would use such technology...

Until it becomes a revolution!

The number of applications of 3D Reconstruction are so numerous, and the potential so incredible, that it really is a matter of when this will be a mainstream revolution.

And because this is one of the hardest topic I ever got to learn, I decided to create an "exploration" about it to guide others who would want to adventure themselves there...

So, here are a few things we'll learn:

Multi-View Stereo

The first part of the course will teach Multi-View Stereo which is basically 3D reconstruction from multiple images.

3+ View Reconstruction (and 360° reconstruction)

How 2-View Reconstruction Works, and how to extend it to 3+View Reconstruction (Warning: it's not as simple as doing multiple projections, because a lot of overlap might happen)

3D Reconstruction from multiple images

Why you shouldn't use every image of a multi-view dataset, and how to decide which one to pick.

Multi-View Disparity

A 3 Step-Technique to reconstruct 3D objects from multiple images using disparity estimation and image rectification.

Multi-View Stereo Workshop 🧑🏻‍💻

How to implement a 3D Reconstruction algorithm from multiple images (and generate a 360° Point Cloud).

MVS Software

5 Examples of 3D Reconstruction Software (free and paid) on the market (and how they're built)

And many more, including:

The subtle difference between Multi-View Stereo, 3D Reconstruction, Structure From Motion, and Visual SLAM... How to deal with overlaps in your outputs... 3 Applications of 3D Reconstruction  in the industry...Where to find multi-view stereo vision datasets;  7 different types of topologies you should consider before running any 3D Reconstruction project... How reducing the number of images used can improve your final output... The quickest known way to run a state of the art Deep Stereo Vision algorithm on a notebook...

And perhaps, the best of all:
How to combine several reconstructed point clouds into a single one, and "polish" your 3D reconstruction output!

Structure From Motion

The second part we'll explore is called Structure From Motion. It's what's behind Apple's object capture software. Let's see:

The 2 Types of SfM

The 2 Types of Structure From Motion Algorithms (and which one every company is going after)

The 8-Point Algorithm

A special look into the 8-Point Algorithm; helping recover the relative positions of 2 images from feature matching

Essential & Fundamental Matrices

Why it's so important to calculate the Essential & Fundamental Matrices, and how to use them to retrieve camera positions from just images (you'll actually be able to look at 2 images, and know exactly the relative position of the cameras that took the images)

Structure From Motion Workshop 🧑🏻‍💻

A dive into the detailed "from scratch" implementation of a Structure From Motion Algorithm doing 3D Reconstruction (and how to master the essential steps of Structure From Motion in any scenario)

Bundle Adjustment

How to learn about the subtle art of Point Cloud Alignment and use it to remove reprojection errors in your outputs

Plus:

The difference between disparity estimation and feature matching... The 5 Steps to never forget when coding a Structure From Motion algorithm; how to recover epipolar lines and how this information can be used to retrieve a 3D Point Cloud...The mathematical proof that disparity estimation can be reduced to a 1D Stereo Vision search (it's not really helpful, but it's always satisfying to get a mathematical proof)... Exactly what to do with the Essential and Fundamental matrices (and why most engineers learning on their own miss out on the benefits of it... Where to learn about the atrocious maths behind Singular Value Decomposition (and why it's not to be overlooked)...

And a whole lot more...

Abhishek Sreesaila, Senior Machine Learning Engineer at EY

"The dots are connected."

"The topic of 3D reconstruction is a just a complex topic but jeremy distill's down to simple concepts with a portfolio project. All of this is made possible in a few hours! It nicely sets you up for more advanced reading if you are interested to go deep in this area. But also:

Every course in Think Autonomous is connected to every other course! This became obvious to me once I took stereo vision, hydranets and neural optimization. The dots are connected, and so slowly but surely you are on the road to engineering skills only few people possess in the industry."

Frequently asked questions

Can I take the course if I already have Stereo Vision skills, but not the course?

No.

Unless you're able to... solve a quiz!

We have developed a quiz, and if you can get 80% good answers or more, then it would prove you have the skills required.
 
💯 Take the quiz here: https://stereoquiz.thinkautonomous.ai
If you manage to solve it, then the 3D Reconstruction course will open its doors just for you, and you'll have an express ticket to join it!

How long is the exploration and what do I get after it?

The exploration will last about 2 hours. As for all my courses, I will provide ways for you to explore on your own, and you'll have the strong foundations for this.

After completing this course, you'll have skills in 3D Reconstruction, Multi-View Stereo, Deep Disparity Estimation, and Structure From Motion. These skills are needed to work on the most advanced Computer Vision startups like Apple or Tesla.

Do I get lifetime access?

Yes, all courses come with lifetime access, and we make sure to always keep our courses up to date.

  • €159 or 2 monthly payments of €79

3D Reconstruction Exploration

  • Closed

A Deep Exploration of 3D Reconstruction. Advanced Engineers only.

* only accessible to Stereo Vision Engineers, purchase the Stereo Vision course first.