Learning New Dimensions of Human Visual Similarity using Synthetic Data
Supplementary Material
Submission Number: 3743
- Dataset Examples (Section 3 & Figure 1)
- Nearest Neighbor Visualizations (Section 6 & Figure 8)
- Inversion Visualizations (Section 6 & Figure 9)
- Supplementary Material PDF
Dataset Examples
Examples of triplets from our dataset.
- DreamSim-All Metrics Disagreement
- DreamSim-LPIPS Disagreement
- DreamSim-DINO Disagreement
- DreamSim-CLIP Disagreement
- Random
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
Humans
DreamSim
CLIP
DINO
LPIPS
- DreamSim-All Metrics Disagreement
- DreamSim-LPIPS Disagreement
- DreamSim-DINO Disagreement
- DreamSim-CLIP Disagreement
- Random
Nearest Neighbor Visualizations
Nearest neighbor searches in ImageNet-R and COCO
Ours = DreamSim.
| LPIPS | DISTS | OpenCLIP | DINO | Ours | |
|---|---|---|---|---|---|
| Input |  |
|
|
|
|
 |
 |
 |
 |
 |
|
| Nearest Neighbors |  |
 |
 |
 |
 |
 |
 |
 |
 |
 |
| LPIPS | DISTS | OpenCLIP | DINO | Ours | |
|---|---|---|---|---|---|
| Input |  |
|
|
|
|
 |
 |
 |
 |
 |
|
| Nearest Neighbors |  |
 |
 |
 |
 |
 |
 |
 |
 |
 |
| LPIPS | DISTS | OpenCLIP | DINO | Ours | |
|---|---|---|---|---|---|
| Input |  |
|
|
|
|
 |
 |
 |
 |
 |
|
| Nearest Neighbors |  |
 |
 |
 |
 |
 |
 |
 |
 |
 |
| LPIPS | DISTS | OpenCLIP | DINO | Ours | |
|---|---|---|---|---|---|
| Input |  |
|
|
|
|
 |
 |
 |
 |
 |
|
| Nearest Neighbors |  |
 |
 |
 |
 |
 |
 |
 |
 |
 |
| LPIPS | DISTS | OpenCLIP | DINO | Ours | |
|---|---|---|---|---|---|
| Input |  |
|
|
|
|
 |
 |
 |
 |
 |
|
| Nearest Neighbors |  |
 |
 |
 |
 |
 |
 |
 |
 |
 |
| LPIPS | DISTS | OpenCLIP | DINO | Ours | |
|---|---|---|---|---|---|
| Input |  |
|
|
|
|
 |
 |
 |
 |
 |
|
| Nearest Neighbors |  |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Inversion Visualizations
Inversion visualization using optimization, deep image prior, and guided diffusion.
Ours = DreamSim.
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |
| Target | DINO | OpenCLIP | Ensemble | Ours | |
|---|---|---|---|---|---|
| Run 1 |  |
 |
 |
 |
 |
| Run 2 | |
 |
 |
 |
 |
| Run 3 | |
 |
 |
 |
 |
| Run 4 | |
 |
 |
 |
 |
| Run 5 | |
 |
 |
 |
 |