Image-to-Sound Retrieval on SoundingEarth 10m GSD 1.0 (test)
0.45R@100GeoCLAP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP, Text=true, Address=true2023.09 | 0.45 | 143 | |
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP, Text=true, Address=false2023.09 | 0.441 | 152 | |
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP, Text=false, Address=true2023.09 | 0.441 | 153 | |
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP, Text=false, Address=false2023.09 | 0.396 | 199 | |
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP-frozen, Text=false, Address=false2023.09 | 0.384 | 274 | |
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP-frozen, Text=true, Address=false2023.09 | 0.34 | 369 | |
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP-frozen, Text=true, Address=true2023.09 | 0.337 | 378 | |
| GeoCLAPImage Encoder=SATMAE, Text-Audio Encoder=L-CLAP-frozen, Text=false, Address=true2023.09 | 0.311 | 453 | |
| BaselineImage Encoder=ResNet18, Text-Audio Encoder=ResNet18, Text=false, Address=false2023.09 | 0.289 | 620 |