Task 1. ChatGPT¶

Prompt: can you give me fun riddles related to 3d computer vision!

Output:

I capture the world, but not in 2D, I need two eyes, just like you see. With depth I measure near and far, What am I, in vision, by par?

(Answer: Stereo Vision / Stereo Camera)

I’m not a painter, yet I splat, A cloud of Gaussians—imagine that! I render scenes, both sharp and grand, Which new method do I command?

(Answer: 3D Gaussian Splatting)

One camera can’t guess, but I can infer, By moving around, things become clearer. From parallax shifts, the depths I unveil, What’s this technique, without fail?

(Answer: Structure from Motion)

I’m a map, but not of roads, I hold values at every node. Brighter means closer, darker far away, What am I in CV, can you say?

(Answer: Depth Map)

Triangles are my building blocks, I’m used in games and vision talks. With vertices and edges, surfaces I weave, What 3D structure do I achieve?

(Answer: Mesh)

I’m invisible but guide your scan, From Kinect to iPhone, I lend a hand. I bounce back quickly, telling you “where,” What sensing method puts me there?

(Answer: Time-of-Flight / LiDAR)

Not a magician, but I can project, From world to image, I must respect. A matrix helps me map the view, Which fundamental tool is true?

(Answer: Camera Projection Matrix)

Task 2. 3D shape images¶

source: self created image + google VEO + gif converted