Summary
Investigate integrating 3D, spherical, and potentially 4D image datasets into the web pattern generator as stimulus templates or built-in examples. May work best as a standalone companion application rather than embedded in the main tool.
Candidate Datasets
| Dataset |
Type |
Images/Scenes |
Key Features |
Format |
License |
| SYNS (Southampton-York) |
Spherical |
~100 scenes |
LiDAR range + HDR imagery, panoramic stereo pairs |
Co-registered HDR + LiDAR |
Research |
| Laval Outdoor HDR |
Spherical |
205 panoramas |
Full-HDR (22 f-stops) outdoor |
Equirectangular, multi-exposure |
Research |
| Poly Haven |
Spherical |
500+ nature HDRIs |
360° HDR panoramas, time/weather metadata |
EXR, HDR |
CC0 (Public Domain) |
| UPenn Natural Image Database |
Planar |
~4000 scenes |
Calibrated Nikon D70, baboon habitat |
Luminance, RGB, LMS |
Research |
| 3D60 |
Spherical |
Multi-modal stereo |
RGB, Depth maps, Normal maps |
PNG, EXR |
Research |
| Leader360V |
Spherical Video |
10K+ videos |
360° video with instance segmentation (198 categories) |
ERP 2048x1024, 30fps |
Research |
| WildScenes |
LiDAR + 2D |
9,306 images + 12,148 submaps |
Semantic annotations in 2D and 3D |
LiDAR + images, 6-DoF |
Research |
Technical Considerations
- Need to handle standard formats: equirectangular projection, HDR/EXR, LiDAR point clouds
- Conversion pipeline from spherical images to arena pixel patterns
- Resolution mapping (high-res source → low-res LED arena)
- May need downsampling and quantization (to GS2/GS16 levels)
Approach
This may work best as a standalone companion application rather than embedded in the main Pattern Editor:
- Upload spherical image → preview on 3D arena model → extract arena-resolution pattern
- Could use existing arena-geometry.js for coordinate mapping
- Consider both static frames and video/temporal sequences
Dependencies
Priority
Future enhancement / exploratory
Summary
Investigate integrating 3D, spherical, and potentially 4D image datasets into the web pattern generator as stimulus templates or built-in examples. May work best as a standalone companion application rather than embedded in the main tool.
Candidate Datasets
Technical Considerations
Approach
This may work best as a standalone companion application rather than embedded in the main Pattern Editor:
Dependencies
Priority
Future enhancement / exploratory