Overview
Gaussian splatting is a 3D scene reconstruction technique that represents environments as collections of 3D Gaussians. Each Gaussian stores position, covariance, color, and opacity, and the renderer composites them into photorealistic views from any camera angle. Unlike mesh-based representations, Gaussian splats capture fine details like foliage, reflections, and semi-transparent surfaces. In Avala, you can upload Gaussian Splat files to a dataset and annotate them directly in the viewer — placing 3D cuboids, assigning classification labels, and organizing annotations in a scene hierarchy.Viewer Layout
The viewer is organized into five regions:| Region | Position | Width / Height | Content |
|---|---|---|---|
| 3D Canvas | Center | Fills remaining space | Interactive Gaussian Splat rendering with camera controls |
| Scene Hierarchy | Left panel | 280px | Tree view of all objects and annotations in the scene |
| Properties | Right panel | 300px | Editable properties for the selected annotation (position, dimensions, rotation, label) |
| Toolbar | Bottom center | 56px height | Annotation tools and mode selection |
| Mode Toggle | Top center | — | Switch between navigation and annotation modes, plus undo/redo controls |
| Statistics | Bottom right | — | Real-time render statistics (frame rate, splat count, GPU memory) |
Navigation
Explore the reconstructed scene using standard 3D camera controls:| Control | Action |
|---|---|
| Left-click + drag | Orbit the camera around the focal point |
| Right-click + drag | Pan the camera laterally |
| Scroll wheel | Zoom in and out |
| Middle-click + drag | Fly-through navigation for moving the focal point |
WebGPU Rendering
The viewer uses a WebGPU-accelerated rendering pipeline optimized for real-time Gaussian splatting.Pipeline Architecture
| Component | Purpose |
|---|---|
| GPU radix sort | Sorts splats by depth on the GPU for correct back-to-front compositing |
| Buffer pooling | Reuses GPU buffer allocations to reduce memory allocation overhead |
| Render bundle caching | Pre-records GPU command sequences that do not change between frames |
| Pipeline precompilation | Compiles WGSL shaders asynchronously during loading to avoid rendering stalls |
Browser Support
| Browser | Version | Status |
|---|---|---|
| Chrome | 113+ | Enabled by default |
| Edge | 113+ | Enabled by default |
| Firefox | Nightly | Requires dom.webgpu.enabled flag |
| Safari | Technology Preview | Requires feature flag |
Annotation
Place annotations directly in the 3D Gaussian Splat scene using the toolbar at the bottom of the viewer.3D Cuboids
Draw three-dimensional bounding boxes around objects in the scene:- Select the Cuboid tool from the toolbar
- Click in the scene to place the cuboid center
- Adjust position, dimensions, and rotation using the gizmo handles
- Assign a label from your project’s label taxonomy
Classification Labels
Apply scene-level or object-level classification labels:- Scene classification — Label the entire scene (e.g., indoor, outdoor, parking lot)
- Object classification — Label individual cuboid annotations with object classes and attributes
Scene Hierarchy
The left panel displays a tree view of all annotations, organized by label class. Use the hierarchy to:- Select annotations by clicking their name
- Toggle annotation visibility
- Reorder and group annotations
- Inspect annotation properties without navigating to them in the 3D view
Edit History
Full undo/redo support is available via the top-center controls or keyboard shortcuts:| Shortcut | Action |
|---|---|
Ctrl+Z / Cmd+Z | Undo |
Ctrl+Shift+Z / Cmd+Shift+Z | Redo |
Use Cases
3D Scene Understanding
Annotate objects in reconstructed indoor and outdoor environments for 3D object detection and scene graph generation.
Novel View Synthesis
Label objects in Gaussian Splat scenes to create training data for view synthesis models that need semantic awareness.
Spatial AI Training Data
Build labeled datasets for embodied AI agents that need to understand and navigate 3D environments.
Digital Twin Annotation
Annotate reconstructed real-world environments for digital twin applications, facility management, and simulation.
Next Steps
3D Cuboid Tool
Detailed guide for placing and adjusting 3D bounding boxes.
3D Point Cloud Viewer
GPU-accelerated point cloud visualization with six rendering modes.
Data Types
Overview of all supported data modalities and their capabilities.
Multi-Sensor Viewer
Synchronized playback of camera, LiDAR, radar, and IMU data.