Frontory

Gaussian Splatting

Frontory technology map

Graphics · 4 stages · 28 units in the full series · 28 published in English

Series goal

Understand 3D Gaussian data and the projection and compositing paths, and design a fault-tolerant viewer using small data with verified provenance.

Prerequisites

Three.js cameras and projection, alpha blending

Practice approach and scope

Learning units are not calendar days. Read the explanation and tasks, check the completion criteria, then move on. Use each example as a starting point: change variables and observe the results. Built-in examples and tasks requiring a separate environment are identified separately. Opening an example does not complete a unit.

Existing unit IDs and notes are preserved. The original 01, 02, and 03 examples occupy positions 1, 14, and 28 in the expanded sequence, so the number in the address may differ from the learning order.

Only published English units are linked below. The full series has 28 units.

Learning path

Stage 1 · Foundations of Representation and Projection

Stage 2 · File Validation and Viewer Integration

Stage 3 · Visibility and Quality Experiments

Stage 4 · A Measurable GS Exhibition Viewer

Capstone project and deliverables

A local GS viewer with a fixed camera path, loading states, a selection policy, and a quality comparison report

Keep the implementation, run instructions, baseline and changed conditions, failure cases, and recovery steps together. Do not report unmeasured performance or tests you have not run as results.

Official references