Assignment 0

Author: Po-Ting Ko

Task 1: ChatGPT

Prompt:

Explain Structure from Motion as if it were a detective solving a mystery.

ChatGPT Answer:

Imagine a detective investigating a crime scene. The detective does not have a 3D map of the room, but they have several photographs taken from different positions.

First, the detective looks for clues that appear in multiple photographs. For example, a corner of a table might appear in three different photos. These matching points help the detective recognize that they are looking at the same real-world object.

Next, the detective compares how these points appear to move between the photographs. If the camera moved, nearby objects will appear to shift more than distant objects. From these changes, the detective can estimate where the camera was and which direction it was facing when each photo was taken.

Once the camera positions are estimated, the detective can use geometry to determine the 3D position of the objects. By combining information from multiple photographs, the system creates many 3D points, eventually forming a 3D representation of the scene.

The process can be summarized as:

Take multiple photos → find matching features → estimate camera movement → calculate 3D points → create a 3D scene

So, like a detective piecing together clues, Structure from Motion reconstructs a 3D scene by analyzing how the same visual features change across multiple images.

Task 2: 3D Shape Image

For this task, I chose the Stanford Bunny, a classic 3D model commonly used in computer graphics research.

Stanford Bunny

Source: Stanford Computer Graphics Laboratory

Task 3: Funny GIF

For this task, I chose a meme GIF from GIPHY.

Funny cat meme GIF

Source: GIPHY